SHRP-A-402

Water Sensitivity:Binder Validation

Todd V. ScholzRonald L. TerrelAbdulla AI-Joaib

Jungju Bea

Strategic Highway Research ProgramNational Research Council

Washington, DC 1994

SHRP-A-402Contract A-003AISBN 0-309-05810-4

Product no.: 1001

Program Manager: Edward T. HarriganProject Manager: Rita B. LeahyProduction Area Secretary: Juliet NarsiahProduction Editor: Margaret S. Milhous

May 1994

key words:

water-sensitivitymoisture-damage

strippingasphalt concrete mixesbinder validation

Environmental Conditioning System (ECS)

Strategic Highway Research ProgramNational Research Council2101 Constitution Avenue N.W.

Washington, DC 20418

(202) 334-3774

The publication of this report does not necessarily indicate approval or endorsement by the National Academyof Sciences, the United States Government, or the American Association of State Highway and Transportation

Officials or its member states of the findings, opinions, conclusions, or recommendations either inferred orspecifically expressed herein.

©1994 National Academy of Sciences

1.SM/NAP/594

Acknowledgments

The work reported herein has been conducted as part of project A-003A of the StrategicHighway Research Program (SHRP). SHRP is a unit of the National Research Council thatwas authorized by Section 128 of the Surface Transportation and Uniform RelocationAssistance Act of 1987.

The typing by Teresa Culver, the graphics by Gail Barnes, and the overall guidance andreport review at Oregon State University are gratefully acknowledged.

°°o

111

Contents

Acknowledgments ............................................ iii

List of Figures ............................................. vii

List of Tables .............................................. ix

Abstract .................................................. 1

Executive Summary ........................................... 3

1 Introduction ........................................... 7

1.1 Background ....................................... 7

1.2 Purpose ......................................... 7

2 Hypotheses of A-002A and A-003B ................ : ............ 9

2.1 A-002A ...................................... . . . 9

2.2 A-003B ......................................... 12

3 Experiment Design ....................................... 17

3.1 Variables Considered ................................. 17

3.2 Materials ........................................ 22

3.2.1 Aggregates and Their Properties ...................... 223.2.2 Asphalts and Their Properties ....................... 22

3.3 Specimen Preparation ................................. 22

3.4 Testing Methods .................................... 30

3.4.10SU ECS Test ................................ 30

V

3.4.2 OSU Wheel-Tracking Test ............................ 383.4.3 SWK/UN Wheel-Tracking Test ......................... 43

4 Results .................................................... 47

4.1 ECS Test Program ....................................... 47

4.2 OSU Wheel-Tracking Program .............................. 55

4.3 SWK/UN Wheel-Tracking Program ........................... 55

4.4 University of Nevada (Reno) Net Adsorption Test (NAT/UNR)Program .............................................. 64

5 Analysis of Results ........................................... 67

5.1 Statistical Analysis ....................................... 67

5.1.1 ECS Test Results ................................... 67

5.1.2 OSU Wheel-Tracking Test Results ...................... 725.1.3 SWK/UN Wheel-Tracking Test Results ................... 725.1.4 NAT/UNR Results .................................. 76

5.2 Performance Ranking ..................................... 76

5.2.1 Aggregates ....................................... 765.2.2 Asphalts ......................................... 785.2.3 Mixes ........................................... 78

5.3 Comparison of A-002A and A-003B Results ..................... 84

5.4 Discussion Regarding Specifications .......................... 87

6 Conclusions and Recommendations ................................ 89

6.1 Conclusions ........................................... 89

6.2 Recommendations ....................................... 90

7 References ................................................. 91

Appendix ATest Data ....................................................... 93

vi

List of Figures

Figure 2.1 Relationship between weight percent SEC fraction II andtan delta for SHRP asphalts (after Robertson 1991) ............... 11

Figure 3.1 Schematic of the specimen preparation process ................. 31

Figure 3.2 Asphalt-aggregate mixer used at OSU ....................... 32

Figure 3.3 Rolling wheel compactor used at OSU ...................... 32

Figure 3.4 The compacted slab .................................. 33

Figure 3.5 Layout of specimens cut from the slab ...................... 34

Figure 3.6 Schematic of the ECS ................................. 35

Figure 3.7 The OSU wheel tracker ............................... 39

Figure 3.8 Schematic of the OSU wheel tracker ........................ 39

Figure 3.9 Deformation measurement positions in the OSU wheel tracker ........ 42

Figure 3.10 Schematic of the SWK/UN wheel tracker ..................... 44

Figure 3.11 Schematic of test specimen fixed into the mold of SWK/UNwheel tracker ...................................... 45

Figure 4.1 ECS test results for the RC aggregate ....................... 52

Figure 4.2 ECS test results for the RD aggregate ....................... 52

Figure 4.3 ECS test results for the RH aggregate ....................... 53

Figure 4.4 ECS test results for the RJ aggregate ....................... 53

Figure 4.5 The effect of ECS conditioning on water-permeabilityof RC mixes ...................................... 56

vii

Figure 4.6 OSU wheel-tracking test results for the RC aggregate ............. 60

Figure 4.7 OSU wheel-tracking test results for the RD aggregate ............. 60

Figure 4.8 OSU wheel-tracking test results for the RH aggregate ............. 61

Figure 4.9 OSU wheel-tracking test results for the RI aggregate .............. 61

Figure 4.10 NAT results ...................................... 66

Figure 5.1 Ranking of 32 mixes after one and three cycles ................. 85

Figure 5.2 Ranking of 32 mixes by NAT and ECS test ................... 85

°°.

Vlll

List of Tables

Table 2.1 Rank of High Temperature Permanent Deformation by SEC-Tan Delta .. 10

Table 2.2 Rank of Moisture-Damage Resistance by Infraredof Functional Group Analysis ............................... 13

Table 2.3 Initial and Net Adsorption of Asphalt on Aggregate ............... 14

Table 3.1 Experiment Design for Validation of Binder Specifications--Water-Sensitivity ....................................... 18

Table 3.2 Experiment Design of Water-Sensitivity Testing Program ............ 19

Table 3.3 Job-Mix Formula Used for the Validation Study .................. 21

Table 3.4 Aggregate Characteristics .................................. 23

Table 3.5 Asphalt Characteristics .................................... 25

Table 3.6 Summary of Specimen Preparation Procedure for SHRP A-003AValidation Study--Water-Sensitivity .......................... 29

Table 3.7 Summary of ECS Test Procedure ............................ 36

Table 3.8 Summary of OSU Wheel-Tracking Test Procedure ................ 40

Table 3.9 Summary of SWK/UN Wheel-Tracking Procedure ................. 46

Table 4.1 Summary of ECS Test Data for RC Mixes ...................... 48

Table 4.2 Summary of ECS Test Data for RD Mixes ...................... 49

Table 4.3 Summary of ECS Test Data for RH Mixes ...................... 50

Table 4.4 Summary of ECS Test Data for RJ Mixes ...................... 51

Table 4.5 Summary of Mixes Tested in the OSU Wheel-Tracking Program ...... 57

ix

Table 4.6 Rut Depths for the OSU Wheel-Tracking Program ................ 59

Table 4.7 SWK/UN Wheel-Tracking Test Results ........................ 62

Table 4.8 NAT Program ......................................... 65

Table 5.1 Variables Considered in the Analyses of the ECS Test Results ........ 69

Table 5.2 An Overview of the ECS Statistical Analyses ................... 70

Table 5.3 GLM Analysis of the ECS Results for Asphalt and Aggregate Type .... 71

Table 5.4 GLM Analysis of the OSU Wheel-Tracking Test Results ............ 73

Table 5.5 Bayesian Survival Analysis of the SWK/UN Test Results ........... 75

Table 5.6 GLM Analysis of the NAT Results for Asphalt and Aggregate Type .... 77

Table 5.7 Performance Ranking of Aggregates Based on ECS Test ............ 77

Table 5.8 Performance Ranking of Aggregates (OSU Wheel-Tracking Program) ... 79

Table 5.9 Performance Ranking of Aggregates (SWK/UN Wheel-Tracking Program) 79

Table 5.10 Performance Ranking of Aggregates (NAT/UNR Test Program) ....... 79

Table 5.11 Performance Ranking of Asphalts Based on ECS Test .............. 80

Table 5.12 Performance Ranking of Asphalts (OSU Wheel-Tracking Program) ..... 80

Table 5.13 Performance Ranking of Asphalts (SWK/UN Wheel-Tracking Program) 81

Table 5.14 Performance Ranking of Asphalts (NAT/UNR Program) ............ 81

Table 5.15 Ranking of 32 Mixes after Each ECS Cycle ..................... 82

Table 5.16 Summary of Aggregate Rankings ............................ 86

Table 5.17 Summary of Asphalt Rankings .............................. 86

X

Abstract

This study was part of the Strategic Highway Research Program (SHRP) A-003A contractundertaken as an interim evaluation of the hypotheses concerning the influence of binderproperties on the moisture susceptibility of asphalt-aggregate mixes.

Thirty-two mixes using eight asphalts and four aggregates from the Materials ReferenceLibrary (MRL) were used to fabricate roller-compacted slabs from which specimens weresawed or cored. These mixes were tested by four procedures: (1) EnvironmentalConditioning System (ECS), (2) Oregon State University (OSU) Wheel Tracker, (3) SWKPavement Engineering/University of Nottingham (SWK/UN) Wheel Tracker, and (4) NetAdsorption Test (NAT).

Because the water-conditioning and testing procedures were different, it was not appropriateto make direct comparisons of results or to use the results for ranking materials. The ECSwas the only test to compare dry- to wet-conditioned mixes, while the two wheel trackerstested only wet-conditioned mixes. The NAT was not directly comparable because itconsiders only adhesion. The two rutting tests provided results that were similar to theA-002A hypothesis, but did not confirm the ECS results. It was concluded that, for water-sensitivity, asphalts or aggregates could not be ranked alone but that combinations or pairswere more appropriate because of strong interactions.

Executive Summary

The original proposal by the Strategic Highway Research Program (SHRP) A-003Acontractor included a major task to evaluate the water-sensitivity of asphalt paving mixes.The testing included a laboratory phase to develop procedures and criteria followed by afield evaluation phase to confirm the performance of various projects that were predicted bylaboratory testing. During the development phase SHRP staff became concerned that datamay not be sufficient or that time-in-service of the field projects may not be adequate.Therefore, an extended program was developed to provide interim data based on acceleratedtests.

The extended test program included evaluation of rutting, thermal and fatigue cracking,aging, and water-sensitivity. Its aim was to perform so-called torture tests that inducedamage or failure in a short time period by simulating loading and environmentalconditioning that would occur on the in-service pavement. For water-sensitivity evaluation,an accelerated rutting test using the Oregon State University (OSU) wheel tracker(Laboratoires des Ponts et Chaus6es [LCPC] rutting tester) was selected as the primarymethod. However, tests on the same mixes were also conducted using the wheel-ruttingtester at SWK Pavement Engineering/University of Nottingham (UK) and theEnvironmental Conditioning System (ECS) developed at OSU. Each test procedure resultsin a different failure mechanism, but all tests can be used to evaluate the water-sensitivityof asphalt-aggregate mixes.

The primary purpose of this portion of the SHRP A-003A program was to validate thehypothesis and preliminary results of asphalt ranking developed by the SHRP A-002Acontractor. Validation of the hypothesis of the SHRP A-003B contractor was alsoconsidered. SHRP A-002A researchers originally suggested a hypothesis of expectedperformance of mixes that is based on asphalt binder properties. They proposed a rankingof the SHRP Materials Reference Library (MRL) asphalts (and to some extent, aggregates)for rutting, aging, fatigue, and water-sensitivity. SHRP A-003B researchers hypothesizedan explanation of the chemical relationship between asphalt, aggregate, and water to defineasphalt-aggregate interactions that are either resistive or sensitive to water, and theydeveloped a Net Adsorption Test (NAT). They proposed rankings of the water-sensitivityof the SHRP MRL aggregates based on adsorption and desorption of asphaltic modelcomponents on aggregates, the net adsorption of asphalts on aggregates, and the percentagechange of adsorption and desorption of selected asphalt models and asphalts on selectedaggregates pretreated with organosilane compounds.

This report covers the testing, analysis of data, and ranking of the MRL materials based onthe ECS, OSU wheel-tracking, and SWK/UN wheel-tracking test results. The ranking ofthe MRL materials is compared with that proposed by the A-002A and A-003B contractors.Also, the NAT, which was performed at University of Nevada-Reno, was used to furthercompare the materials used in the testing program.

An eight-asphalt by four-aggregate (8 x 4) matrix was designed with tests repeated on eightof the mixes. The set of 40 tests (32 mixes plus 8 repeats) was designed to identify thewater-sensitivity of the mixes using either rutting (OSU and SWK/UN wheel-tracking) orreduction in the ECS modulus. The test program provided information to rank the relativeperformance of the eight asphalts and four aggregates, thus enabling a comparison of resultsprovided by the A-002A, A-003A, and A-003B contractors.

Specimens were prepared using a procedure developed at OSU. Batches of each mixturewere heated in ovens and mixed in a converted concrete mixer. Asphalt concrete slabs

approximately 75 cm square by 50 mm thick were prepared using a small tandem wheelroller and a specially constructed mold. After cooling, specimens were sawed and coredfrom the slab for the various tests.

Cores from the slabs were tested in the ECS using the three-hot and one-freeze cycle andcontinuous repeated loading procedures. The ratio of dry and water-conditioned resilientmodulus (ECS-MR) was used to compare the mix.

Small slabs or beams, which were conditioned in a special chamber similar to that used forthe ECS, were sawed from the slabs for testing in the OSU wheel tracker. Tests wereconducted only on conditioned specimens, so the test results were for relative ruttingresistance for each mix; i.e., there was no ratio between wet and dry. Comparisons weremade after 10,000 wheel passes.

Beams were also sawed and shipped to SWK/UN for testing in a wheel tracker. Thisdevice was used to test specimens that were water-conditioned by a combination of soakingand freezing. Wheel-tracking continued on each specimen until it failed or until about500,000 wheel passes were attained.

The results and conclusions of this study can be summarized as follows:

1) Performance ranking of materials by asphalt type or aggregate type alone wasnot feasible for the ECS test results because of the significant interaction

between asphalt and aggregate. In general, aggregate RC and ILl were theworst performers or most water sensitive; aggregate RH was the intermediateperformer; and RD was the best performer or most water resistant.

2) Statistical analysis of ECS test data showed that the test is sensitive toaggregate and asphalt type and can discriminate between moisture-sensitiveand moisture-resistant mixes.

4

3) The OSU wheel-tracking test results indicated that the ILl aggregate was arelatively good performer, the RC aggregate was a poor performer, and theRD and RH aggregates were intermediate performers in terms of rutresistance.

4) The SWK/UN wheel-tracking test results indicated that the RC and RDaggregates were good performers (with practically no difference between thetwo), the RH aggregate was an intermediate performer, and the RJ aggregatewas a poor performer. The significant differences between the results of the

two test methods may possibly be attributed to the significant differences intesting methods, test apparatus, specimen size, or specimen environmentduring testing. However, the results of the SWK/UN wheel-tracking testgenerally validate the predictions proposed by the NAT (A-003B) but thoseof the OSU wheel-tracking test do not. Thus, the OSU wheel-tracking testmay not be appropriate for evaluating the water-sensitivity of aggregate typesunless a wet/dry ratio can be evaluated.

5) Asphalt performance ranking of both wheel-tracking tests showed goodcorrelation with aged asphalt viscosity data. The Size Exclusion

Chromatography (SEC)-tan delta test proposed by A-002A appears to predictadequately the rutting potential of asphalt types as evidenced by closeagreement with the asphalt rankings from the OSU wheel-tracking andSWK/UN wheel-tracking tests. It is interesting to note that near perfectagreement exists between A-002A predictions and SWK/UN results.

6) Predictions of the water-sensitivity of the binder as proposed by the A-002AFourier transform infrared (FTIR) test and the A-003B NAT showed little orno correlation to wheel-tracking tests on the mix. Since very good toexcellent correlation exists between the wheel-tracking tests and the A-002Apredictions for permanent deformation, it appears that the FTIR test and theNAT are poor indicators of the moisture sensitivity of the binder.

Further observation of the results of the tests may be helpful in interpreting them. Therange of mixes (32 combinations) was not as wide as might be expected because the mixdesigns (aggregate gradation, asphalt content) in the laboratory phase were the same, incontrast to those in the field validation phase, which included a much greater variety of mixtypes. Only the ECS test program measured water-sensitivity, i.e., comparison of wet-conditioned to original dry mixes. The OSU and SWK/UN wheel trackers tested water-conditioned specimens, but the results were not compared to dry test results; thus, the testsonly compared similarly conditioned mixes. The NAT is an indicator of potential foradhesion (stripping) only and does not consider the cohesion-loss aspect of water damage.In the final analysis, conclusions must be drawn only after careful consideration of thevariables.

The results of ECS testing showed that the ECS can distinguish between good and poorperforming mixes with respect to moisture-damage. The results of this test program are notsufficient to establish limits for specifications. The results of the 32 mixes tested in this

program were in a relatively narrow range. The testing program included only MRLmaterials with no direct link to field performance. Therefore, the consideration of

specifications will be deferred until completion of field validation testing.

6

1

Introduction

1.1 Background

The original proposal by the Strategic Highway Research Program (SHRP) A-003Acontractor included a major task to evaluate the water-sensitivity of asphalt paving mixes.The testing included a laboratory phase to develop procedures and criteria, followed by afield evaluation phase to confirm the performance of various projects that were predicted bylaboratory testing. During the laboratory phase, SHRP staff became concerned that datamay not be sufficient or that time-in-service of the field projects may not be adequate.Therefore, an extended program was developed to provide interim data based on acceleratedtests. The extended test program included the evaluation of rutting, thermal cracking,aging, and fatigue and water-sensitivity over a range of asphalt and aggregate types. Itsaim was to use so-called torture tests that induce damage or failure in a short time period,by simulating loading and environmental conditioning that would occur on the in-servicepavement.

For water-sensitivity, an accelerated rutting test using the LCPC rutting tester (OSU wheeltracker) was selected as the primary method of evaluation. However, tests on the samemixes were also conducted using the wheel-rutting tester at SWK PavementEngineering/University of Nottingham (UK) (SWK/UN wheel tracker) and theEnvironmental Conditioning System (ECS) developed at OSU. Each test procedure resultsin a different failure mechanism, but all tests can be used to evaluate the water-sensitivityof asphalt-aggregate mixes.

1.2 Purpose

The primary purpose of this portion of the SHRP A-003A program was to validate thehypothesis and preliminary results of asphalt ranking developed by the SHRP A-002Acontractor. Validation of the hypothesis of the SHRP A-003B contractor (AuburnUniversity) was also given consideration. The SHRP A-002A researchers originallypresented a hypothesis of expected performance of mixes based on asphalt binder

7

properties. They proposed a tentative ranking of the SHRP Materials Reference Library(MRL) asphalts (and to some extent, aggregates) for rutting, aging, thermal cracking,fatigue, and water-sensitivity. However, based upon later work the hypothesis and rankingswere rescinded. SHRP A-003B researchers have hypothesized an explanation of thechemical relationship among asphalt, aggregate, and water to describe and defineasphalt-aggregate interactions that are either resistive or sensitive to water. They haveproposed rankings of the water-sensitivity of the SHRP MRL aggregates based onadsorption and desorption of asphaltic model components on aggregates, the net adsorptionof asphalts on aggregates, and the percentage change of adsorption and desorption ofselected asphalt models and asphalts on selected aggregates pretreated with organosilanecompounds.

This report covers the testing, analysis of data, and performance ranking of the MRLmaterials based on the ECS and the OSU and SWK/UN wheel-tracking test results. The

ranking of the MRL materials is compared with that proposed by the A-002A and A-003Bcontractors.

2

Hypotheses of A-002A and A-003B

Since the primary purpose of the Strategic Highway Research Program (SHRP) A-003Awork was the validation of the A-002A and A-003B hypotheses for water-sensitivity, thischapter presents reviews of both the A-002A hypothesis prepared in March, 1991 (Robertson1991) and the A-003B hypothesis (Curtis et al. 1993).

2.1 A-002A

The SHRP A-002A contractor (Western Research Institute) was commissioned to developpredictions for asphalt-aggregate-mixture performance based on the properties of the binder.Mix-performance measures included fatigue, permanent deformation (rutting), aging, thermalcracking, and water-sensitivity (in terms of loss of adhesion). Only the predictions forwater-sensitivity and permanent deformation are considered in this report, permanentdeformation predictions are included since the A-003A water studies used rutting tests as partof the validation effort.

The ranking of asphalts for permanent deformation are shown in Table 2.1 (Robertson 1991).This ranking is based on preparative Size Exclusion Chromatography (SEC) fraction I toSEC fraction II ratios that show a strong correlation with viscoelastic properties of the binderas shown in Figure 2.1. Note that this ranking is based on asphalts that have notexperienced long-term aging. The SEC fraction I is the weight of the nonfluorescentcomponents in the asphalt whereas the SEC Fraction II is the weight of the fluorescentcomponents. The nonfluorescent components assemble into an elastic matrix while the

fluorescent components form the dispersing phase for the matrix. This dispersing phase doesnot appear to self-assemble at moderate to high temperatures and therefore is primarilycomposed of a viscous material. The SEC fraction I to SEC fraction II ratio provides ameasure of the total of the system.

9

Table 2.1. Rank of High Temperature Permanent Deformation by SEC-Tan Delta a

Asphalt Type Resistance

AAM-1 ExcellentAAK-1AAE

AAS-1 Very GoodAAH-1AAD-1AAB-1AAWolAAJ

AAAol GoodAAN

AAX FairAAF°IAAC-1

AAZ PoorAAV

AAG-1 Little or No ResistanceABD

a After Robertson 1991.

10

11

Several studies have demonstrated that loss of adhesion via moisture-damage is primarilyassociated with the aggregate (Curtis et al. 1991; Robertson 1991). The A-002A contractorfeels that classification of the moisture-damage susceptibility from the chemistry of the binderalone is probably a minor effect at best (Robertson 1991). However, A-002A formulated theclassification shown in Table 2.2, which is based on the carbonyl content (with emphasis onthe free acid content) as determined by FTIR. Note that aging affects the asphaltsdifferently.

2.2 A-003B

The SHRP A-003B contractor (Auburn University) was charged with describing and definingasphalt-aggregate interactions that are resistive or sensitive to water. This effort examinedthree specific areas: (1) evaluation of the specific chemistry of asphalt adsorption ontoaggregate using model species that are representative of polar functional group types presentin asphalts; (2) evaluation of the compatibility of various asphalt-aggregate pairs and theirrespective sensitivity to water; and (3) determination of the effect that aggregates treated withsaline compounds of differing chemistries have on asphalt-aggregate interactions and water-sensitivity (Curtis et al. 1991).

The A-003B contractor concluded that the adsorptive behavior of asphalt and asphalt modelcomponents on aggregates is highly specific and particularly influenced by the aggregatesurface chemistry; the chemistry of the asphalt binder has less influence. Tests onaggregates pretreated with saline compounds led to the same conclusion. Experimentsevaluating adsorption and aqueous desorption of asphalt model components on aggregatesconclusively showed that polar compounds had different affinities for adsorption for differentaggregates. The amount and ease by which the polar compounds were removed from theaggregate surface in the presence of water was found to be dependent on the aggregatechemistry as well as the pH and heat history of the particular system.

Net adsorption tests (NATs) were used to investigate the compatibility and water-sensitivityof asphalt-aggregate pairs and clearly showed that the adsorption behavior of asphalt onaggregate was controlled by the aggregate chemistry. The A-003B researchers found thatsubstantial differences in adsorption and aqueous desorption behavior existed amongaggregates but that small and generally insignificant differences existed among asphalts; thatis, the differences in adsorption and desorption behavior of one particular asphalt incombination with various aggregates were far in excess of those of one particular aggregatein combination with various asphalts. Table 2.3 shows the net adsorption data obtained bythe A-003B contractor during the development of the NAT procedure. The initial amount ofasphalt adsorbed before introduction of water gives an indication of the affinity a particularasphalt has for a given aggregate. The net adsorption, or the amount remaining on thesurface of the aggregate after aqueous desorption, is an indication of the moisture sensitivityof the asphalt-aggregate pair (Curtis et al. 1991). The results of the NAT (i.e., ranking of

12

Table 2.2. Rank of Moisture-damage Resistance by Infrared of Functional GroupAnalysisa

New Material Aged Material

Asphalt Resistance Asphalt Resistance

AAF-1 Good (No orderestablished) AAB-1 Good (in orderas shown)AAB-1 AAM-1AAM-1 AAC-1AAAol AAF-1

AAD-1 Intermediate-Good AAD-1 Intermediate-GoodAAK-1

AAG-1 Fair-Poor

ABD Poor AAG-I PoorABD

a After Robertson1991.

13

-_1 o_

• _ d _ _ _ <_ _ _

._

14

aggregates via net adsorption) are used herein to compare the results of the A-002A andA-003A contractors.

Tests evaluate the effect that saline-treated aggregate has on the asphalt-aggregate interactionand water-sensitivity showed enhanced adsorption and retention of asphalt model compoundsand asphalts in specific cases and no change or decreased adsorption and increased water-sensitivity in others. Pretreatment of the aggregate surface (i.e., modification of the surfacechemistry) created an asphalt chemistry that was more specific in both its adsorption anddesorption behavior than that created by untreated aggregate.

The NAT was used by the A-003B contractors on only a few of the Materials ReferenceLibrary (MRL) aggregate-asphalt combinations. Late in the research program, the A-003Acontractors determined that additional NAT results would be beneficial. Accordingly, asubcontract to the University of Nevada-Reno was initiated to test the 32 combinations (eightasphalts x four aggregates) used in the validation experiment. The results of the NAT testare addressed later.

15

3

Experiment Design

The experiment design developed for the Strategic Highway Research Program (SHRP)A-003A task D.2.e work is shown in Table 3.1. The table shows the coding scheme ofeach mix. The first two digits are the aggregate code (i.e., RC and RJ codes are 00 and 11,respectively); the last three digits are the asphalt code (i.e., AAA-1 and AAG-1 codes are000 and 101, respectively). Originally only eight mixes were chosen to be replicated;however, all the 32 mixes were actually replicated. The Environmental ConditioningSystem (ECS) validation phase was divided into two tasks: 1) Laboratory validation, usingthe ECS, and 2) Field validation, using two wheel-tracking systems Laboratories des Pontset Chaus6es (LCPC) and SWK Pavement Engineering (SWK).

As indicated, an eight-asphalt by four-aggregate (8 x 4) matrix was designed for this workwith tests repeated on eight of the mixes. (Duplication of tests in the ECS and OregonState University (OSU) wheel-tracking programs exceeded that which was designed asdiscussed in further detail in Chapter 4.) The set of 40 tests (32 mixes plus eight repeats)was primarily designed to identify the water-sensitivity of the mixes using either rutting(OSU and SWK/UN wheel-tracking) or reduction in modulus (ECS) as the objectivefunction. The test program provided information to rank the relative performance of theeight asphalts and four aggregates, thus enabling a comparison of results provided by theA-002A, A-003A, and A-003B contractors. The following sections provide detailsregarding the experiment design including the variables considered, the materials employed,the specimen preparation procedure, and the test procedures used to carry out the work.

3.1 Variables Considered

The testing program consisted of eight asphalt types and four aggregate types. The ECStest program variables considered for this phase of the SHRP project are shown in Table3.2 and discussed below:

17

Table 3.1. Experiment Design for Validation of Binder Specifications--Water-

Sensitivity

Mix Mix MRL MRL Required

Number Code Aggregate Asphalt Replicate

1 00000 RC AAA- I RC & AAA- 1

2 10000 AAB-13 01000 AAC-14 11000 AAD- 1

5 00100 AAF-1

6 10100 AAG-17 01100 AAK-I RC & AAK-1

8 11100 AAM-1

9 00010 RD AAA-1I0 10010 AAB-111 01010 AAC-112 11010 AAD-1 RD & AAD-1

13 00110 AAF-I14 10110 AAG-1 RD & AAG-115 01110 AAK-116 11110 AAM-1

17 00001 RH AAA-118 10001 AAB-I19 01001 AAC-120 11001 AAD-1 RH & AAD-1

21 00101 AAF-122 10101 AAG-1 RH & AAG-123 01101 AAK-124 11101 AAM-1

25 00011 RJ AAA-1 RJ & AAA-126 10011 AAB-127 01011 AAC-128 11011 AAD- 1

29 00111 AAF- 1

30 10111 AAG-131 01111 AAK-1 RJ & AAK-132 11111 AAM-I

18

Table 3.2. Experiment Design of Water-Sensitivity Testing Program

Level of Treatment Number

Controlled ] [ ofVariable 1 2 3 Levels

Asphalt type

* Grade 2 Low 2 Medium High 8• Content Optimum 1

Aggregate type

• Stripping potential Low 2 Medium High 4• Gradation Medium 1

Compaction

• Air voids (%) 8+1 1

Testing compaction factors

• Test temperature 3 Hot Cycles (60°C) + Freeze Cycle 1

• Load (-18°C) Repeated 1• Time 6 hr 1

Total 32

Complete factorial 32

Replicate 3_.22Total number of samples 64

Response variables:

• Initial ECS modulus

• Initial air-permeability• ECS-M R after each cycle• Water-permeability after each cycle

• Visual evaluation, percentage of retained asphalt coating on the aggregate

19

1) Specimen density (air-voids), mix asphalt content, and gradation of theaggregate were all held as constant as possible.

2) Permeability was used as a measure of the moisture-damage susceptibility.Air-void's range of 8 + 1 percent was used to control the permeability.

3) Temperatures that were applied during conditioning were hot (60°C [140°F])for the first three cycles and freeze (-18°C [-0.4°F]) for the fourth cycle. Theresilient modulus (ECS-MR) 1 test was conducted at 25°C (77°F) after eachcycle.

4) Repeated loading was applied during the first three hot cycles, and staticloading during the freeze cycle.

5) The specimen was preconditioned or saturated with distilled water at 50.8 cm(20 in.) Hg of vacuum for 30 minutes.

6) The duration of each cycle was 6 hours, and each test had three hot cyclesand one freeze cycle.

Response variables:

1) ECS-M R was measured after each conditioning cycle.

2) Permeability was measured after each conditioning cycle to monitor thechange in moisture-damage susceptibility.

3) Visual estimation of the percentage of retained asphalt coating on theaggregate was observed at end of test.

Full factorial experiment design was used (Table 3.2). The order of sample preparation wasrandomized independently for each replicate, and the specimens were selected and testedrandomly.

For OSU and SWK/UN wheel-tracking test programs, the variables considered in theexperiment design included asphalt and aggregate type. Specimen density (air-voids), mixasphalt content, gradation of the aggregate, and test specimen conditioning were all held asconstant as possible: specimen air-voids' contents here held constant at 8+1 percent; themix-asphalt contents were based on the content established by the Hveem Method and aregiven in Table 3.3; the aggregate was of medium gradation (see Table 3.3); and each test

1The resilient modulus obtained in the ECS is termed the ECS-MR to distinguish it from the traditional diametraland triaxial resilient moduli as well as from the dynamic modulus. The ECS-M Ris a triaxial resilient modulus withzero confining stress (i.e., a2 = a3 = 0) conducted on a 102 mm (4 in.) diameter by 102 mm (-_4 in.) tallasphalt-aggregate mixture test specimen (Terrel and A1-Swailmi 1992).

20

Table 3.3. Job-Mix Formula for the Validation Study

Percent Passing

Sieve Size RC RD RH RJ

1 in. 100 I00 100 100

3/4 in. 95 95 95 95

1/2 in. 80 80 80 80

3/8 in. 68 68 68 68

#4 48 48 48 48

#8 35 35 35 35

#16 25 25 25 25

#30 17 17 17 17

#50 12 12 12 12

#100 8 8 8 8

#200 5.5 5.5 5.5 5.5

Asphalt content by 6.25 4.5 5.2 5.0weight of aggregate,

Asphalt content by total 5.9 4.3 4.9 4.8weight of mix,

21

program employed a conditioning procedure that remained the same for all specimens tested(each method is described in further detail below).

3.2 Materials

The materials used in this study included eight asphalts and four aggregates from the SHRPMaterials Reference Library (MRL). The following paragraphs provide details of thesematerials.

3.2.1 Aggregates and Their Properties

Two limestones (RC and RD) and two siliceous aggregates (RH and RJ) were used for thisresearch effort. Table 3.4 summarizes the properties of the aggregates available at the time

this report was written. Note that the RC limestone aggregate had high water absorptionand centrifuge kerosene equivalent (CKE) values compared with the other aggregates. RDaggregate showed very low absorption values. In addition, the RC aggregate had a lowbulk specific gravity compared with that of the other aggregates (the gravimetric data forthe RH aggregate were unavailable). In the soundness test, RC aggregate exhibited highvalues of percentage loss of fine and coarse fraction compared with the other aggregates.

3.2.2 Asphalts and Their Properties

Eight asphalts from differing sources (crudes) and having differing grades were used in thisresearch effort. The MRL codes for these asphalts are AAA-1, AAB-1, AAC-1, AAD-1,AAF-1, AAG-1, AAK-1, and AAM-1. Table 3.5 summarizes the properties of theseasphalts. Note the wide range of asphalt viscosities as determined by the traditionalviscosity and penetration tests. These data show that the AAC-1 asphalt is the softest andthe AAK-1 asphalt is the hardest of the asphalts, based on original asphalt viscosity at 60°C(140°F).

3.3 Specimen Preparation

Specimen preparation for this research effort was accomplished by means of rolling wheelcompaction. Table 3.6 gives a brief description of the procedure.

22

Table 3.4. Aggregate Characteristics

MRL Code RC RD RH RJ

Major element oxide

SiO2 5.58 (11.79) 16.68 (14.84) 75.91 75.4 (63.98)TiO2 0.06 (0.18) 0.13 (0.21) 0.46 0.15 (0.41)A1203 1.18 (1.46) 3.31 (1.95) 10.68 12.88 (14.6)Fe203 0.76 (0.89) 1.2 (0.96) 4.83 2.01 (4.54)CaO 48.92 (35.04) 38.8 (33.71) 1.84 1.73 (6.09)MgO 2.35 (11.76) 3.47 (11.43) 2.28 0.39 (1.52)Na20 0.17 (0.21) 0.12 (0.08) 2.76 3.4 (1.67)

K20 0.18 (0.51) 1.56 (2) 0.74 3.31 (3.31)Sulfur Trioxide (0.48) (0.34) (0.1)Phosphorus pentoxide (< 0.01) (< 0.01) (0.11)Manganicoxide (0.03) (0.02) (0.13)

LOI 40.62(37.64) 33.96(34.45) 2.41 I.13 (3.54)

Composition% Limestone, Limestone, Micaceous Sandstone,47.4I00 53.3 Sandstone, Granite,28.4

Limestone, 71.3 Misc.,23.7

26.8 Misc.,11.2 Basalt,0.4

Arenaceous Granite, 10.9Limestone, Chert, 6.619.7

Porosity (ASTM D-4404)

Average pore diameter (mxl0 -6) (0.0611) (0.0111) (0.0151)Total pore area (m2/g) (2.548) (1.465) (1.888)

Mercury porosimetry data

Pore size A Pore vol., cc/g Pore vol., cc/g Pore vol., cc/g Pore vol., cc/g> 300 0.0099 0.0013 0.0128 0.0026500-3000 0.1085 0.0301 0.0905 0.0071< 500 0.0045 0.0003 0.0023 0.0002Total volume 0.12 0.03 0.11 0.01

pH 9.7 9.8 8.6 9.6Los Angeles abrasion (AASHTOT-96)Wear % (39.1) (23.4) (29.5)Water Absorption (AASHTOT-84, T-85)

Absorption % (3.7) (0.3) (0.7)

Specific gravity (AASHTOT-84, T-85)

Bulk (2.536) (2.704) (2.550) (2.625)Saturated surface dry (2.595) (2.717) (2.646)Apparent (2.682) (2.739) (2.741 ) (2.68)

Continued on next page.23

Table 3.4. Aggregate Characteristics (continued)I I

MRL Code RC RD RH RJ

BET Surface area, m2/g 2.9 .72 2.74 1.32Rootare-Prenzlow .84 0.14 0.53 0.05

Surface area (mZ/g) 7.9 (4.8) 23.5 (18.1) 92.1 96.2 (99.2)Acid insolubles (%) 8.1 (2.4) 5.1 (1.9) 9.7 6.3 (4.1)

Water insolubles (%) -6.1 @pH9.82 - 13.6@pH9.87 -20.5@pH8.27 -27.5@pH9.45Zeta potential (-23.8) (-20.3) (-49)

CKE (AASHTO T-270)Uncorrected (%) (8.5) (3.8) (1.8)Oil retained (%) (3.9) (2.7) (2.6)

Flakiness index (%) (22.6) (34.7) (9.6)

(Asphalt Institute)Sand equivalent (%) (32) (69) (60)

(AASHTO T-176)

Magnesium soundness(AASHTO T-104)

Loss (%): Fine fraction (6.32) (1.52) (1.29)Loss (%): Coarse fraction (0.51) (0.04) (0.16)

Polish value (ASTM D-3319)BPN before polish (42) (38) (41)BPN after polish (31) (28) (22)

Data from University of Kentucky; data in parentheses from Southwestern Lab, Inc., Texas.

24

0

0

c_ ',,c:> _ _ o', o ,,_.o.,

"9, _

(,.J _ _ . . .

(,,; _ r,-_ _ _ mr-

"_" ',"-' ,,,-_ t"-,, ,--, ,,-_ "--_ _"_ _ c_ ,,--_

._ _ ,

_I _ i _° __ _°"°

"_ _ _E_ _

O"--""--" <'<_Z

25

26

0

n_

"_,

0

_ _ N_,_ . g_ .

_ _ _o___ .

_ _ _ _oo._ _ _ o__.=

L

,=

,..._,_., '_ ,_ _.

°_ °_ _ _

['_ _ _

27

_o,_ i oo o__ ,o _ I

I_ _ _'_'1 _:'m_i_ I _" "-; _:

ii "7' ' _,

• t_l_

:__I _° _ _I._._1

28

Table 3.6. Summary of Specimen Preparation Procedure for SHRP A-003A

Validation Study--Water-Sensitivity

Step Description

1 Calculate the quantity of materials (asphalt and aggregate) needed based on the volume of the mold,

the theoretical maximum (Rice) specific gravity of the mix, and the desired percent air-voids.Batch weights ranged between 125 to 132 kg (275 and 290 lb) at an air-void content of8 _+1 percent.

2 Prepare the asphalt and aggregate for mixing.

3 Heat the materials to the mixing temperature for the asphalt (170_+20 cSt). Mixing temperaturesranged between 137 ° and 160°C (279 ° and 320°F).

4 Mix the asphalt and aggregate for 4 minutes in a conventional concrete mixer fitted with infrared

propane burners and preheated to the mixing temperature for the asphalt.

5 Age the mix at 135°C (275°F) in a forced-draft oven for 4 hours stirring the mix every hour torepresent the amount of aging that occurs in the mixing plant.

6 Assemble and preheat the compaction mold using infrared heat lamps.

7 Place the mix in the compaction mold and level it using a rake while avoiding segregation of themix.

8 Compact the mix when it reaches the compaction temperature using a rolling wheel compactor untilthe desired density is obtained. This is determined by the thickness of the specimen (the onlyvolumetric dimension that can be varied during compaction for a set width and length of slab).Steel channels with depth equal to the thickness of the specimen prevent overcompaction of the mix.Compaction temperatures (based on 630+20 cSt) ranged between 112 ° and 133°C (234 ° and271 °F).

9 Allow the compacted mix to cool to room temperature ( = 15 hr).

10 Disassemble the mold and remove the slab. Dry cut (saw) beams for the OSU and SWK/UN wheeltrackers. Dry cut cores for the ECS.

29

The specimen preparation process is shown schematically in Figure 3.1. The mixer usedconsisted of a conventional concrete mixer modified to include infrared propane heaters

(see Figure 3.2) to preheat the mixer bowl prior to mixing as well as to reduce heat lossduring the mixing process. The preheated and preweighed aggregate is added to the mixerfollowed by the asphalt. The mix, typically 125 to 132 kg (275 to 290 lb), is mixed in onebatch. After mixing, the asphalt-aggregate mix is placed in a forced-draft oven set to135°C (275°F) and short-term aged for 4 hours to simulate the amount of aging that occursin a batch or drum-dryer plant. The mix is stirred once each hour to promote uniformaging. At the completion of the aging process, the mix is placed in the mold andcompacted to a predetermined density using a small steel wheel compactor with tandemrollers (e.g., a roller for compacting sidewalks and bike paths). The compactor used atOSU (Figure 3.3) weighed approximately 3260 lb (_1480 kg). The compacted slab(Figure 3.4) was then allowed to cool overnight (15 hr) after which beam specimens weresawn and core specimens were drilled from the slab (Figure 3.5). The beams were sawnand the cores were drilled without water to prevent errors in density and void analysis aswell as in initial air-permeability tests. For air-permeability tests and bulk specific gravitythe specimen must be dry; water in voids can hinder the air-flow through the specimen,thus giving wrong air-flow numbers and air-permeability results.

3.4 Testing Methods

Each test program (ECS, OSU wheel-tracking, and SWKdUN wheel-tracking) employedspecimen conditioning in its test procedure that subjected the specimen to water damagefollowed by measurement of rutting (OSU and SWK/UN wheel trackers) or reduction inmodulus (ECS). This section briefly describes these procedures; detailed test methods arefound elsewhere (A1-Joaib 1993).

3.4.10SUECS Test

The test procedure employed in the ECS program consisted of inducing and monitoringwater damage to 10-cm (4 in.) diameter by 10-cm (4 in.) high asphalt concrete cores. Theprocedure is briefly described in Table 3.7. Figure 3.6 shows a schematic of the equipmentused in the ECS test program. The system consists of three components:

1) Environmental chamber with controlled temperature and humidity.

2) Fluid conditioning system that is essentially a constant headpermeameter with the fluid medium being either water or conditionedair (dry or moist) and a thermocouple controller with fourthermocouples installed on the ECS panel to monitor the temperature of

3O

_ o _

0CJ °_

c_

o

v

31

\

Fig. 3.2. Asphalt-aggregate mixer used at OSU

Fig. 3.3. Rolling wheel compactor used at OSU

32

Fig. 3.4. The compacted slab

33

{5

ECS Cores

[_ Waste

(a) Plan View

L (0.81 mm)

\i .......... /TI I I _ I I I I I

......... mmI I I I I I I I I I

L

0.17 m

(b) Elevation View

Fig. 3.5. Layout of specimens cut from the slab

34

mm

8"-,'v_ :8,v_ ... __g

-a_-.0, Fo

_'' qg:N _

I :m _

I,lilt-" *"_,, me-

• _ "_ .

e- _:_ _-__ ,,_ _..---_EE _-o_=C j_ r--..LI.I.I?,,_0 0

35

Table 3.7. Summary of ECS Test Procedure

Step Description

1 Preparetest specimens as describedin Table 3.6. The specimen size is 4 inch (102 ram) diameterby 4 inch (102 ram)height.

2 Determine the geometric and gravimetric quantities of the core.

3 Place a silicone seal around the circumferenceof the core with a 6 inch membraneand allow thesilicone cement to cure overnight(_- 24 h).

4 Mount the core specimen in the ECS load frame and determine the air-permeabilityof the core atvarious flow levels.

5 Determine the unconditioned (dry) resilient modulus.

6 Apply 20 inch (508 mr.) Hg vacuum for 10 minutes.

7 Wet the core by pulling distilled water through the specimen for 30 minutes using a 20 inch(508 ram)Hg vacuum.

8 Determine the unconditioned water-permeabilityof the core.

9 Heat the wet specimen to 60"C (140*F) for 6 hours and apply axial repeatedloading of 18 psi(_ 124 kPa). This constitutesa hot cycle.

10 Cool the wet specimen to 25"C (77"F) for 2 hours and measure the water-permeability and resilientmodulus.

11 Repeatsteps 8 and 9 for two (2) more hot cycles.

12 Cool the wet specimen to -18"C (0*F) for 6 hours. This constitutes a freeze cycle.

13 Heat the specimen to 25"C (77"F) for 2 hours and measure the water-permeability and resilientmodulus.

14 Split the specimen and assess the percentageof stripping.

15 Plot water-permeability and resilient modulus ratios (conditionedto unconditioned) versusconditioningcycle.

36

the water before entering the specimen, after entering the specimen, and insidea dummy specimen in the chamber.

3) Computer-controlled loading and data acquisition system to monitor the axialresilient modulus (ECS-MR) of the test specimen.

The ECS test is carried out to assess quantitatively the effect water has on the stiffness andpermeability of an asphalt-aggregate mix. Prior to testing, gravimetric data (specificgravities) are obtained for the core specimen. The specimen is then encapsulated in a latexmembrane with silicone. In the test, the dry (unconditioned) ECS-M R and air-permeabilityare determined prior to introduction of water. The specimen is then wetted by flowingdistilled water through the specimen under the action of a negative pressure relative toatmospheric pressure (i.e., vacuum) for 30 minutes. Upon completion of the wettingprocess, the water-permeability of the specimen is determined. The specimen is thensubjected to thermal conditioning cycles, consisting of three hot cycles by heating thespecimen to 60°C (1400F) and one freeze cycle by cooling the specimen to -18°C (0°F).The duration of each thermal cycle is 6 hours. Between cycles the specimen is brought to25°C (770F) and tested to determine the conditioned water-permeability and ECS-MR, thusmonitoring the effect water has on these properties as a function of the type and amount ofenvironmental conditioning.

Important test parameters in the ECS test include the following:

1) All material property testing (modulus and permeability) is conducted at atemperature of 250C (770F). Also, only one specimen setup is needed,eliminating errors caused by handling when modulus or permeability test isconducted.

2) The modulus test is a triaxial test with a zero confining pressure (i.e., a2 = tr3= 0), herein referred to as an axial resilient modulus test. The load (i.e.,deviator stress), in the form of a true haversian waveform having a duration of0.1 s followed by a dwell time of 0.9 s, is targeted to be 40 psi (_- 275 kPa).Sufficient conditioning loads with magnitude equal to the target load areapplied to the specimen prior to obtaining modulus data to ensure constantplastic deformation at the time data are obtained.

3) The test specimen is loaded automatically by a computer program that uses aclosed-loop proportional-derivative (PD) feedback algorithm in conjunctionwith additional hardware to drive a servovalve-air-piston system and acquireload and deformation data. Such a system helps to minimize user errors.

4) Repeated loading of 18 psi (_ 124 kPa) is applied throughout the hot cycles tosimulate traffic loading.

37

3.4. 20SU Wheel-Tracking Test

The test procedure employed in the OSU wheel-tracking program consisted of inducing waterdamage to beams of asphalt-aggregate mixes having dimensions of approximately 19 incheslong by 6.5 inches wide by 4 inches deep (_-483 mm x 165 mm x 102 mm) andmonitoring the rut-depth developed in the OSU wheel tracker. Figure 3.7 shows aphotograph of the OSU wheel tracker and Figure 3.8 is a schematic of this equipment. Theprocedure is briefly described in Table 3.8. Note that dry (unconditioned) specimens werenot tested since the purpose of the program was to rank the materials (to validate the A-002Apredictions), which required that all specimens be tested in the same manner.

The OSU wheel-tracking test, a torture test, is performed to obtain a relative measure of therutting resistance among asphalt-aggregate mixes after the mixes have been subjected towater-conditioning. Prior to testing, gravimetric data are obtained for the beam specimenfollowed by subjecting the specimen to water-conditioning. The conditioning procedureemployed to induce water damage in the beams used for the OSU wheel-tracking program isessentially the same as that for the ECS test program, except for the following minordifferences:

1) The wetting procedure for the wheel-tracking test program employs a slightlyhigher vacuum level and a significantly longer wetting time than that for theECS test to achieve the target saturation level of 60 to 80 percent in the largerbeam specimens, although a few beams did not reach the this level.

2) The duration of some of the conditioning cycles is longer in the OSU wheel-tracking test procedure than in the ECS test procedure because of schedulingconstraints of some of the equipment used for thermal conditioning.

3) The order of conditioning cycles for the wheel-tracking program is slightlydifferent from that of the ECS test program, also because of schedulingconstraints of some of the equipment used for thermal conditioning.

Once the beam specimen has undergone water and thermal conditioning, the specimen iswrapped in plastic (e.g., Saran wrap) to prevent moisture loss. The specimen is then placedin a mold for subsequent testing in the OSU wheel tracker. Thin expanded foam sheets areplaced between the specimen and the mold walls to prevent movement under the action of therolling wheel. A Teflon sheet approximately 3 mm (1/8 in.) thick and having the same plandimensions as the specimen is placed under the specimen to minimize the friction thatdevelops between the specimen and base platen during the test. The mold is then placed inthe wheel tracker and brought to the test temperature of 40°C (104°F). The plastic wrap isremoved from the top surface of the specimen so as to prevent the plastic from being pickedup by the pneumatic tire.

38

)panel _ Pneumatic

I_ tire

ll _.. Specimen

Fig. 3.7. The OSU wheel tracker

Tire TireRail

Stub axleStub axle

Swivel

Swivel

91atform

Jack Jack

Telescopic arm

Fig. 3.8. Schematic of the OSU wheel tracker

39

Table 3.8. Summary of OSU Wheel-Tracking Test Procedure

Step Description

1 Prepare test specimens as described in Table 3.6.

2 Determine the gravimetdc quantities of the beam.

3 Place a circumferential silicone cement seal around the beam at midheight and allow the siliconecement to cure overnight (=24 hr).

4 Apply 20-inch Hg (508) Hg vacuum for 10 minutes.

5 Wet the beam specimen by pulling distilled water through the specimen under a 23-inch (584 ram)vacuum level for up to 2 hours or until a degree of saturation of at least 60 is obtained.

6 Subject the wet beam specimen to wet thermal conditioning cycles as follows:

• Heat the specimen to 60°C (140°F) in a distilled water bath for 6 hours.• Cool the specimen to 25°C (77°F) in a distilled water bath for 10 hours.• Heat the specimen to 60°C (140°F) in a distilled water bath for 6 hours.• Cool the specimen to -20°C (-4°F) in a distilled water bath for 8 hours.• Heat the specimen to 60°C (140°F) in a distilled water bath for 10 hours.

• Cool the specimen to 25°C (77°F) in a distilled water bath for 10 hours.

7 Wrap the specimen in plastic (e.g., Saran wrap) to retain moisture in the specimen during therutting phase.

8 Place the conditioned beam specimen in the rutting tester and heat the specimen to 40°C (104°F).

9 Perform the OSU wheel-tracking (rutting) test on the conditioned beam specimen until 10,000 wheelpasses have elapsed, taking rut-depth measurements at O, 200, 500, 1,000, 2,000, 5,000, and10,000 wheel passes.

10 Plot rut-depth versus wheel passes.

11 Core the rutted beam specimen along the wheel track to obtain cores for stripping evaluation. Splitthe cores and assess the percentage of stripping.

40

After the specimen reaches the test temperature, determined by a thermocouple probeinserted in a hole drilled in the specimen, preconditioning wheel loads of 50 wheel passes at92 psi (_ 635 kPa) are applied to the beam specimen to eliminate the high plasticdeformations characteristic of asphalt-aggregate mixes at the onset of loading. Afterpreconditioning, the load is removed and measurements are obtained to establish the base-linespecimen surface profile. Figure 3.9 shows the 15 positions in which surface prof'llemeasurements are obtained. These measurements are obtained electronically (i.e., viacomputer), using a displacement transducer specifically designed for these measurements.Note that the measurement positions are concentrated near the center of the specimen alongits longitudinal axis to avoid measurement of high plastic deformations that occur in theregion where the rolling wheel slows down, stops, and finally reverses direction (i.e., at theends of the wheel travel).

The wheel load is then reapplied and increased to 100 psi (-_ 690 kPa). Testing commencesby applying up to 10,000 wheel passes or until failure occurs (as established by a sudden andsignificant increase in plastic deformation). At intervals of 100, 200, 500, 1,000, 2,000, and5,000 wheel passes, the load is temporarily removed so that surface profile measurementscan be obtained. After 10,000 wheel passes (or when loading is terminated because ofspecimen failure), the final surface profile is determined. From these data the rut-depth isdetermined as a function of the number of wheel passes.

Important parameters in the OSU wheel-tracking test include the following:

1) Wheel: pressurized pneumatic tire, 40.6 cm (16 in.) diameter by 10.2 cm (4.0in.) width; smooth tread with 83 mm (3.25 in.) width.

2) Preconditioning load: 50 wheel passes at 92 psi (_ 635 kPa) actual contactpressure.

3) Test load: 10,000 wheel passes at 100 psi (_-.690 kPa) actual contact pressure(1,600-lb. load with tire tread contact area of 16 in2).

4) Load frequency: 60 cycles per minute (120 wheel passes per min.).

5) Test specimen temperature: 40°C (IlM°F).

6) Confinement: base provides reaction to the load; initially unconfined on sides,partially confined as specimen deforms.

7) Environment: conditioned specimen wrapped in plastic (except for the topsurface) tested in air at 40°C (104°F).

41

5.9 in. 3.5 in. 3.5 in. 5.9 in.(150 ram) (90 ram) (90 ram) (150 ram)

iii]iiii!iiiiiiiiiiiiiiiiiii]::iii?iii:iiiiii]iiiiiiiiiiii:i:i:i:i:]:i:i:i:i:i:i:!:i3!:]:i:i:i:i:i:!:i:i:!:i:i:::iiiii:ii!ii iii"iii!i!

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::!:i:i:!:i:i:i:i:!:i:i:i:i:i:!_-:-!:i::!::::i :::iiiii]itiiiiiiiiiiii:.':.':::':._i:.iiiiiiiiiiiii_iiiiiiiiiiiiiiiiiiiiiiiii.O.iiiiiiiii!i!i!iii!i]i.O.iiii!iii!_---/i iii_iiii: : : :_ _:::]_::_._;:.::_:_:_:_:[_:];_._:_:::_:::_:_:::::_:._.F:.[._.:._.F[._.:._.;.F:.F_._._._._": ; _; >]';".'_%z-__z_;;;_::]_:_:i

::::::::::::::::::: ::::::: :::::::::i:i:i:i:i:i::.==========================================================..::::::::::::_::::

:'[-Z':-_':':':-:':':-:':':':"_ _iiI.';'Z'Z':'F_':':. '.FF:':':':':: :'____[.Q;: Z; i : ::: _[ [ : Z[ ; ZZ_: : : ] Z_;;__ ; [ : 2_Z_[_I-_-:-Z::: =============================================================================================================:::::::::::::::::::::::

Measurement Region of highRut position plastic deformation

(both ends)outline a) Plan view

i

Wheel travel

2 •17.7 in. (450 mm) l

/_ ..... J

b) Elevation view

Fig. 3.9. Deformation measurement positions in the OSU wheel tracker

42

3.4. 3 SWK/UN Wheel-Tracking Test

The test procedure employed in the SWK/UN wheel-tracking program consisted of inducingwater damage to beams of asphalt-aggregate mixes having dimensions of approximately 12inches long by 3.5 inches wide by 1 inch deep (305 mm x 90 mm x 25 mm) andmonitoring the specimen surface deformation developed in the SWK/UN wheel tracker.Figure 3.10 provides a schematic of this equipment. The SWK/UN wheel-tracking test,also a torture test, is carried out to obtain a relative measure of the rutting resistance amongasphalt-aggregate mixes after the mixes have been subjected to water-conditioning. Prior totesting, gravimetric data are obtained for the beam specimen. The specimen is then bondedin the mold for subsequent conditioning and testing (Figure 3.11). The specimen is thensubjected to water-conditioning. Note that significant differences exist between the wheel-tracking test conditioning procedures at OSU and SWK/UN (compare Tables 3.8 and 3.9).In particular, note that the duration and number of cycles are quite different. Thetemperatures for conditioning and testing are the same however.

Once the specimen has been water-conditioned, it is placed in the wheel tracker andbrought to the temperature of 40°C (104°F). The specimen is submerged in a water bathduring the SWK/UN wheel-tracking test. The specimen is then loaded with the wheel andtesting commences. The test continues until failure (as determined by a sudden andsignificant increase in plastic deformation of the specimen) or, alternately, until seven daysof loading (_500,000 wheel passes) have occurred. Every 20 wheel passes deformationdata are obtained that consist of measurements of the vertical position of the wheel viaLVDTs and a strip-chart recorder.

Key parameters regarding the SWK/UN wheel-tracking test include the following:

1) Wheel: steel wheel, 20.16 cm (7.9 in.) diameter by 50.4 mm (2 in.) width.

2) Preconditioning load: none.

3) Test load: up to 500,000 wheel passes at 41 pounds (181 N).

4) Load frequency: 25 cycles per minute (50 wheel passes per minute).

5) Test specimen temperature: 40°C (104°F).

6) Confinement: confined on all sides throughout the test; the base providesreaction to the load.

7) Environment: conditioned specimen tested submerged in water at 40°C(104°F).

43

_3n_ _ wheel _ I_f_a_

r 2._lmm_azol=, I / _ _ _ stl_Otnl_todmt

/,__, / I _-_ I,. •

7iiiii iililil_A\ ; " t "1_"

weir "F_'_mm _ _ ..... I_ Motor-operated' iI lI.,...__ ram '_ _".... .-: ....... I 7" crank,--_1 ....' I--,---305_ ........... , 24rpm

c _ I --_-,,"l- ...... sr=y_....::\1 _ '= .......-....._,_

JL __ ___ °]_

Fig. 3.10. Schematic of the SWK/UN wheel tracker

44

STEEL CASING SPECIMEN

SILVER FOIL EPOXY RESIN

Fig. 3.11. Schematic of test specimen fixed into the mold of SWK/UN wheel tracker

45

Table 3.9. Summary of SWK/UN Wheel-Tracking Procedure

Step Description

1 Prepare specimens (at OSU) as described in Table 3.6. Ship to the University of Nottingham.

2 Saw the specimen to size and determine gravimetric quantities for the beam specimen.

3 Condition the beam specimen as follows:

• Soak specimen in water at 140*F (60"C) for 120 hours.• Freeze specimen in air at -200C (-4*F) for 24 hours.• Soak specimen in water at 600C (1400F) for 24 hours.• Soak specimen in water at 40"C (104*F) for 2 hours.

5 Perform the SWK/UN wheel-tracking test on the conditioned specimen until failure or, alternatively

if no failure occurs, after seven days of testing (=500,000 wheel passes). The specimen is

submerged in 40"C (104*F) water during the test. Deformation measurements, as determined bythe vertical position of the wheel, are recorded every 20 wheel passes.

6 Report time-to-failure in hours.

46

4

Results

This chapter presents the results of the validation efforts for water-sensitivity. The resultsobtained in the Environmental Conditioning Systems (ECS) and Oregon State University(OSU) wheel-tracking programs conducted at OSU as well as those obtained in the SWKPavement Engineering/University of Nottingham (SWK/UN) wheel-tracking programconducted at the University of Nottingham (UK) are included.

4.1 ECS Test Program

The mixes tested in the ECS program are summarized in Tables 4.1 through 4.4. Asindicated, two tests were conducted on each mix, thus exceeding the minimum requirementof eight repeated tests. Tables 4.1 through 4.4 summarize the ECS test program data byaggregate: RC, RD, RH, and RJ, respectively. They include average data for each mix, andall data are included in Appendix A.

The test results for the ECS test program are shown graphically in Figures 4.1 through 4.4.Note that each data point represents the average of two tests and that the line connectingthe data points represents the trend in retained resilient modulus (ECS-M_) ratio as afunction of conditioning level (each 6-hour block represents a conditioning cycle with thefirst three cycles being hot cycles and the last cycle being the freeze cycle); that is, theplots show the ratio of conditioned resilient modulus to unconditioned resilient modulus forseveral conditioning cycles. Thus, the ECS-M Rratio provides an indication of the amountof water damage sustained by the test specimen with the dry (and unconditioned) ECS-M Rbeing the datum.

Figure 4.1 shows the effect of ECS conditioning on all RC mix combinations. Thepreconditioning stage and first conditioning cycle cause the asphalt to soften and mix toexhibit cohesion loss. Cohesion loss is the first step of water damage, and cohesion losstends to enhance or accelerate the adhesion-loss mechanism. After the first cycle, mixes

47

Table 4.1. Summary of ECS Tests Data for RC Mixesi

Air ECS Retained Water Retained StrippingAsphalt Voids Cycle MR MR Perm. Penn. Rate

Type (%) No. (ksi) Ratio E-3 em/s Ratio (%)

8.7 0 190.0 1.00 4.41 1.008.7 1 184.0 0.97 3.58 0.81

AAA-1 8.7 2 180.0 0.95 2.89 0.668.7 3 172.5 0.91 2.87 0.65

8.7 4 162.5 0.86 2.56 0.58 15.0

9.4 0 252.5 1.00 4.68 1.00

9.4 1 245.5 0.97 3.53 0.76AAB-1 9.4 2 228.0 0.90 2.78 0.59

9.4 3 226.0 0.90 2.76 0.599.4 4 206.5 0.82 2.46 0.53 15.0

9.0 0 305.0 1.00 4.96 1.009.0 1 262.5 0.86 3.69 0.74

AAC-1 9.0 2 255.0 0.84 3.20 0.659.0 3 251.5 0.82 2.71 0.559.0 4 228.5 0.75 2.28 0.46 20.0

9.0 0 238.0 1.00 1.88 1.009.0 1 202.0 0.85 2.03 1.08

AAD-1 9.0 2 192.5 0.81 1.87 0.999.0 3 186.0 0.78 1.71 0.919.0 4 181.0 0.76 1.64 0.87 10.0

8.7 0 485.0 1.00 5.83 1.008.7 1 468.0 0.96 2.53 0.43

AAF-1 8.7 2 423.0 0.87 2.14 0.378.7 3 385.0 0.79 1.81 0.318.7 4 374.5 0.77 1.63 0.28 20.0

10.3 0 362.5 1.00 8.97 1.0010.3 I 354.0 0.98 4.99 0.56

AAG-1 10.3 2 338.5 0.93 4.12 0.4610.3 3 321.5 0.89 3.47 0.3910.3 4 292.0 0.81 2.27 0.25 20.0

9.3 0 265.0 1.00 7.41 1.009.3 1 238.0 0.90 4.68 0.63

AAK-1 9.3 2 235.8 0.89 4.02 0.54

9.3 3 231.0 0.87 3.64 0.499.3 4 217.8 0.82 3.39 0.46 15.0

10.1 0 255.0 1.00 9.60 1.0010.1 1 245.1 0.96 5.91 0.62

AAM-1 10.1 2 236.0 0.93 4.91 0.5110.1 3 235.5 0.92 4.18 0.4310.1 4 226.1 0.89 4.02 0.42 10.0

ksi = 6,890 kPa

48

Table 4.2. Summary of ECS Tests Data for RD Mixes

Air ECS Retained Water Retained StrippingAsphalt Voids Cycle MR MR Perm. Penn. Rate

Type (_) No. (ksi) Ratio E-3 cm/s Ratio (%)

8.1 0 187.3 1.00 1.92 1.008.1 1 183.3 0.98 3.40 1.77

AAA-1 8.1 2 179.1 0.96 2.98 1.558.1 3 176.4 0.94 2.80 1.468.1 4 174.8 0.93 2.72 1.42 10.0

8.0 0 277.5 1.00 4.81 1.008.0 1 262.5 0.95 4.69 0.98

AAB-1 8.0 2 245.1 0.88 4.13 0.868.0 3 241.7 0.87 3.96 0.828.0 4 234.5 0.85 3.56 0.74 5.0

8.6 0 265.0 1.00 9.93 1.008.6 1 255.0 0.96 7.22 0.73

AAC-1 8.6 2 248.5 0.94 6.75 0.688.6 3 240.2 0.91 6.44 0.65

8.6 4 234.8 0.89 6.44 0.65 5.0

9.0 0 206.5 1.00 7.20 1.009.0 1 201.5 0.98 5.41 0.75

AAD-1 9.0 2 182.9 0.89 4.20 0.589.0 3 174.4 0.84 4.78 0.669.0 4 174.6 0.85 4.73 0.66 10.0

9.7 0 570.0 1.00 4.38 1.009.7 1 547.5 0.96 5.80 1.33

AAF-1 9.7 2 514.8 0.90 5.52 1.269.7 3 498.9 0.88 5.21 1.199.7 4 490.0 0.86 5.04 1.15 10.0

8.2 0 528.0 1.00 1.12 1.008.2 I 491.7 0.93 2.36 2.10

AAG-1 8.2 2 473.5 0.90 2.18 1.948.2 3 464.9 0.88 2.17 1.938.2 4 488.1 0.92 2.14 1.91 15.0

8.4 0 290.0 1.00 2.42 1.008.4 1 274.6 0.95 3.40 1.40

AAK-1 8.4 2 271.1 0.93 3.45 1.438.4 3 170.0 0.93 3.43 1.428.4 4 276.3 0.95 3.43 1.42 5.0

10.3 0 357.5 1.00 1.45 1.0010.3 1 342.8 0.96 3.06 2.11

AAM-1 10.3 2 324.7 0.91 2.55 1.7610.3 3 316.5 0.89 2.79 1.9310.3 4 318.5 0.89 2.81 1.94 5.0

ksi = 6,890 kPa

49

Table 4.3. Summary of ECS Tests Data for RH Mixes

Air ECS Retained Water Retained StrippingAsphalt Voids Cycle MR MR Perm. Perm. Rate

Type ( _ ) No. (ksi) Ratio E-3 cm/s Ratio ( %)

8.0 0 126.5 1.00 5.85 1.008.0 1 119.2 0.94 4.62 0.79

AAA-1 8.0 2 113.7 0.90 4.29 0.738.0 3 120.3 0.95 3.46 0.598.0 4 118.7 0.94 3.78 0.65 7.5

8.3 0 230.0 1.00 0.06 1.008.3 1 226.5 0.98 2.50 45.05

AAB-I 8.3 2 208.5 0.91 2.09 37.668.3 3 212.5 0.92 2.09 37.66

8.3 4 208.5 0.91 1.79 32.25 10.0

6.9 0 230.5 1.00 0.006.9 1 252.0 1.09 0.12 1.00

AAC-1 6.9 2 269.5 1.17 0.09 0.746.9 3 259.5 1.13 0.07 0.606.9 4 259.5 1.13 0.06 0.55 10.0

7.3 0 201.0 1.00 0.00

7.3 1 192.0 0.96 1.43 1.00AAD-1 7.3 2 190.5 0.95 1.88 1.32

7.3 3 185.5 0.92 1.44 1.01

7.3 4 184.0 0.92 1.61 1.13 7.5

7.3 0 564.5 1.00 0.08 1.007.3 1 471.7 0.84 1.41 17.58

AAF-1 7.3 2 431.3 0.76 1.22 15.197.3 3 446.7 0.79 1.16 14.447.3 4 444.0 0.79 1.14 14.25 10.0

6.4 0 625.0 1.00 0.05 1.006.4 1 566.8 0.91 2.33 46.50

AAG-1 6.4 2 555.5 0.89 0.13 2.606.4 3 553.4 0.89 0.09 1.806.4 4 551.4 0.88 0.07 1.30 10.0

8.0 0 364.5 1.00 1.68 1.008.0 1 306.5 0.84 2.65 1.57

AAK-1 8.0 2 301.0 0.83 2.69 1.60

8.0 3 287.5 0.79 2.22 1.328.0 4 284.0 0.78 2.02 1.20 15.0

7.0 0 415.0 1.00 0.007.0 1 346.0 0.83 2.29 1.00

AAM-1 7.0 2 322.3 0.78 0.14 0.067.0 3 332.5 0.80 1.50 0.657.0 4 327.2 0.79 1.44 0.63 10.0

I

ksi = 6,890 kPa

5O

Table 4.4. Summary of ECS Tests Data for RJ Mixes

Air ECS Retained Water Retained Stripping

Asphalt Voids Cycle MR MR Perm. Perm. RateType (%) No. (ksi) Ratio E-3 cm/s Ratio (%)

8.2 0 145.5 1.00 2.09 1.008.2 I 135.4 0.93 1.26 0.60

AAA-1 8.2 2 129.4 0.89 0.94 0.458.2 3 128.5 0.88 0.34 0.168.2 4 126.7 0.87 0.08 0.04 7.5

8.4 0 337.5 1.00 4.54 1.008.4 1 328.8 0.97 1.66 0.37

AAB-1 8.4 2 286.2 0.85 0.54 0.128.4 3 281.7 0.83 O.14 0.038.4 4 273.1 0.81 0.13 0.03 12.5

7.2 0 300.0 1.00 4.29 1.007.2 1 241.5 0.81 3.95 0.92

AAC-1 7.2 2 219.5 0.73 3.04 0.717.2 3 212.3 0.71 2.41 0.567.2 4 209.0 0.70 2.25 0.53 7.5

7.5 0 185.0 1.00 3.74 1.007.5 1 157.7 0.85 1.86 0.50

AAD-1 7.5 2 148.0 0.80 0.31 0.037.5 3 145.5 0.79 0.11 0.037.5 4 138.9 0.75 0.08 0.02 10.0

8.5 0 426.3 1.00 1.87 1.008.5 I 424.0 0.99 0.88 0.47

AAF-1 8.5 2 406.4 0.95 0.71 0.388.5 3 385.2 0.90 0.31 0.178.5 4 355.0 0.83 0.04 0.02 20.0

8.8 0 352.5 1.00 5.85 1.008.8 1 302.6 0.86 2.73 0.47

AAG-1 8.8 2 264.9 0.75 2.35 0.408.8 3 236.6 0.67 2.09 0.368.8 4 240.6 0.68 1.98 0.34 10.0

8.5 0 265.0 1.00 4.22 1.008.5 1 218.6 0.82 3.71 0.88

AAK-1 8.5 2 214.2 0.81 3.34 0.798.5 3 203.1 0.77 3.19 0.768.5 4 213.0 0.80 3.37 0.80 5.0

8.6 0 299.0 1.00 2.43 1.008.6 1 272.8 0.91 2.14 0.88

AAM-1 8.6 2 260.7 0.87 2.01 0.838.6 3 245.9 0.82 1.60 0.668.6 4 234.1 0.78 0.87 0.36 12.5

ksi = 6,890 kPa

51

1.1 -!I I I

AAA-1I I I

_1.0_ I I _AAB-1

o.8 , , i(_ I I I AAF-1(.3 XILl 0.7 I I I AAG-1

I I t -O

0.6 I I I AAK-1

0 1 2 3 4

Cycle No.

Fig. 4.1. ECS test results for the RC aggregate

It

1.1 AAA-1I I II 1 I

o AAB-1".._ 1.0 t I f .en,' I I AAC-1

-4-

AAD-1= 0.9"O -XO I ---'---

I I "4- - _ _ AAF-1' tco I I I(.3 0.8 AAG-1LU I I I X

1 I 1 AAK-1

0.7 I I I -oAAM-1

0 1 2 3 4Cycle No.

Fig. 4.2. ECS test results for the RD aggregate

52

i

12 , - , / ___f "l o ...... ,_,_1.1 -*" ' ' /

" I I I |

1.o_-_-4,. _ _ / +AAC-1

o_1:_ 0.9

1,,o,08 _o-,0.7 AAK-1

0 1 2 3 4Cycle No.

Fig. 4.3. ECS test results for the RH aggregate

1.1 !1I I I

AAA-11 I I _¢o 1.0 .... -.

m AAB-1n," -. I -o

0.9 .__. AAC-1

o ---= . :----- _ AAD-10.8 T" --x

I

03 I AAF-1

ill 0.7 i ---..J.Z_L.L___AAG-1I IlI_ - :I:

0.6 I I I AAK-1

0 1 2 3 4Cycle No.

Fig. 4.4. ECS test results for the RJ aggregate

53

that have good cohesion properties (i.e., do not lose strength after the first cycle) are notaffected by successive ECS conditioning cycles (i.e., good cohesion improves adhesion orhinders the adhesion-loss).

Other mixes that are susceptible to cohesion loss tend to lose substantial strength after thefirst cycle. After the first cycle, mixes that are susceptible to moisture-damage throughadhesion-loss tend to continue losing strength with each conditioning cycle. Figure 4.1shows that after one cycle of ECS conditioning, the different asphalts fall into two groups.Three asphalts (AAK-1, AAD-1, and AAC-1) that are at or below 0.9 ECS-MR ratio arehighly susceptible to moisture-damage and tend to continue to lose strength with each cycle(cohesion loss at first cycle leads to more adhesion-loss). The other asphalts (not affected bythe first cycle) tend to exhibit small and gradual loss of strength with each cycle. MixRC/AAF-1 is an exception to these observations, because some mixes that are not thoroughlywetted (because of low initial permeability) have minimal cohesion loss. However, after firstcycle the permeability increases and leads to further moisture-damage.

The crisscrossing of the curves for the different asphalts emphasizes that ECS results aredependent on the asphalt type for any given aggregate. Also, ECS results showed that thebehavior of the different mixes change with each cycle (i.e., ranking of mixes changes witheach cycle). In the fourth cycle (freeze) all eight mixes lost strength. In the ECS tests pooraggregates tended to disintegrate throughout the freeze cycle--another moisture-damagephenomena. In aggregate processing and sample preparation, RC aggregate disintegrated.Also, RC aggregate is absorptive (i.e., tends to absorb moisture). This absorptive characterenhances the disintegration potential when subjected to freeze cycle.

Figure 4.2 shows the ECS conditioning effects on all RD aggregate mixes. RD mixturecombinations were less susceptible to ECS conditioning. All mixes showed very slowgradual decrease in strength, i.e., good moisture-damage resistance. The freeze cycle didnot have a significant effect on the strength of the mixes, which can be explained by the factthat RD aggregate is nonabsorptive.

Figure 4.3 is a plot of all RH mixes and shows a wide spread of data. After one cycle threeasphalts lost more than 10 percent of ECS-MR ratio (AAF-1, AAK-1, and AAM-1). Theother five mixes showed ECS-MR ratio of 0.9 or better. Each group maintained its set ofmixes after each cycle, and both groups of asphalts continued to lose strength at very slowrates, which emphasizes that the three asphalt mixes that showed ECS-MR ratio below 0.9after one cycle showed cohesion-loss behavior and not much adhesion-loss. The other fiveasphalt mixes that have ECS-MR ratio above 0.9 showed little cohesion and adhesion-loss(i.e., high moisture-damage resistance). Through the freeze cycle, constant strength wasmaintained, i.e., little moisture-damage and aggregate degradation.

Figure 4.4 shows a plot of aggregate RJ results, and the same observations that were made inaggregate RC can be made here. RJ mixes showed significant moisture susceptibility; eachcontinued ECS-MR loss especially after the first cycle. RJ aggregate proved to be a strippingaggregate (SHRP A-003B). All mix combinations showed gradual decrease in strength afterconditioning cycle.

54

Figure 4.5 is an example of water-permeability plots for RC aggregate. Figure 4.5 showsthe change in water-permeability ratio after each conditioning cycle. The water-permeabilitynormally will decrease after each cycle, because repeated loading tends to rearrange anddensify the mix. In a few incidences, the water-permeability increased after the first cycle.This was the case with specimens that were impermeable or that had very low initialpermeability. Mixes with high air-voids (8 + 1 percent) developed low permeability becauseof lack of interconnection within the void structure. However, after one cycle of repeatedloading at 60°C (140°F), the voids tended to become better connected and the permeabilityincreased. RC and RJ mix combinations exhibited about the same loss in water-permeability,with average final water-permeability ratios of about 0.5 and 0.4, respectively.

4.2 OSU Wheel-Tracking Program

As previously mentioned, the number of tests included in the OSU wheel-tracking programexceeded that which was required by the experiment design. Table 4.5 summarizes themixes tested as well as void content data and percent saturation data for each mix. The lastcolumn in Table 4.5 shows the percent stripping for as many of the mixes as were available.Percent saturation on most of the mixes was not in the desired range of 60 to 80 percentbecause of low initial permeability. As indicated, 25 of the 32 mixes were repeated, thusexceeding the minimum requirement of repeating eight of the tests. In retrospect, it probablywould have been more informative to test one dry- and one wet-conditioned beam rather thanduplicate wet beams in order to provide some measure of water-sensitivity.

The OSU wheel-tracking test results are summarized in Table 4.6. Note that an averagevalue for the rut-depth was used where the mix was replicated (i.e., the results tabulated forreplicated mixes is the average of the two tests performed on the mix). Detailed rut-depthdata for each mix is provided in Appendix A. Graphic representations of the data presentedin Table 4.6 are shown in Figures 4.6 through 4.9. It is clear from these plots that theAAA-1 and AAC-1 asphalt mixes performed worst while the AAK-1 and AAM-1 asphaltmixes performed best with respect to rut resistance.

4.3 SWK/UN Wheel-tracking Program

The test results of the SWK/UN wheel-tracking program are shown in Table 4.7. Note thatSWK/UN reported a time-to-failure in hours where failure is defined as a sudden andsignificant increase in plastic deformation. A pass was reported if the specimen did not

55

_ %"== %'--" %"" _ _ %'--=I I I I I I I I

,

56

Table 4.5. Summary of Mixes Tested in the OSU Wheel-Tracking Program

Mix Aggregate Asphalt Mix Sample Percent Percent PercentNumber Type Type Codea ID b Voids Saturation Stripping

1 RC AAA-1 00000 RR0 7.1 33 25.01 AAA-1 00000 RR1 7.8 55 40.02 AAB-I 10000 RR0 6.9 63 5.02 AAB-1 10000 RR1 6.9 73 25.03 AAC-I 01000 RR0 7.7 64 N/A b3 AAC-1 01000 RRI 7.8 59 30.04 AAD-1 11000 RR0 8.0 65 0.04 ADD-1 11000 RR1 7.4 60 30.05 AAF-I 00100 RR0 7.6 92 5.05 AAF-1 00100 RRI 7.7 66 17.56 AAG-1 10100 RR6 7.9 72 0.07 AAK-I 01100 PRO 7.8 79 5.07 AAK-1 01100 RRI 8.9 61 5.08 AAM-1 11100 PRO 7.7 73 0.08 AAM-I 11100 RR1 8.0 47 5.0

9 RD AAA-1 00010 RR2 8.2 52 N/A9 AAA-1 00010 PR3 8.0 60 5.010 AAB-1 10010 RR2 8.7 45 15.010 AAB-1 10010 RR3 8.4 52 17.511 AAC-1 01010 RR2 8.9 40 5.012 AAD-1 11010 RR0 8.4 57 N/A12 AAD-1 11010 RR1 8.6 56 N/A13 AAF-1 00110 RR0 9.0 56 N/A13 AAF-1 00110 RR1 8.6 49 10.014 AAG-1 10110 RR2 8.7 61 5.014 AAG-1 10110 RR3 8.6 61 0.015 AAK-1 01110 RR2 8.1 51 N/A15 AAK-1 01110 PR3 9.0 63 5.016 AAM-1 11110 RR1 8.6 44 N/A

Continued on next page.

57

Table 4.5. Snmmary of Mixes Tested in OSU Wheel-Tracking Program(continued)

Mix Aggregate Asphalt Mix Sample Percent Percent PercentNumber Type Type Code a ID Voids Saturation Stripping

17 RH AAA-1 00001 RR4 8.2 54 0.017 AAA-1 00001 RR5 7.5 63 12.518 AAB-1 10001 RR3 8.8 42 10.019 AAC-1 01001 RR1 6.9 44 7.519 AAC-1 01001 RR3 6.9 32 5.020 AAD-1 11001 RR0 7.6 46 15.020 AAD-1 11001 RR1 7.8 56 5.021 AAF-1 00101 RR0 8.7 40 30.021 AAF-1 00101 RR1 8.5 57 0.022 AAG-1 10101 RR4 8.7 65 45.022 AAG-1 10101 RR5 8.7 61 35.023 AAK-1 01101 RR0 8.7 43 7.523 AAK-1 01101 RR1 8.8 46 7.524 AAM-1 11101 RR0 7.7 71 5.024 AAM-1 11101 RR1 7.7 38 2.5

25 RJ AAA-1 00011 RR2 8.4 53 N/A25 AAA-1 00011 RR3 8.4 55 N/A26 AAB-1 10011 RR2 7.7 80 5.026 AAB-1 10011 RR3 7.7 55 N/A27 AAC-1 01011 RR7 9.0 63 25.028 AAD-1 11011 RR0 7.2 57 7.528 AAD-1 11011 RR1 7.4 66 N/A29 AAF-1 00111 RR0 8.1 57 N/A29 AAF-1 00111 RR1 8.0 41 N/A30 AAG-1 10111 RR4 8.4 53 70.031 AAK-1 01111 RR0 7.2 47 N/A31 AAK-1 01111 RR1 7.1 50 N/A32 AAM-I 11111 RR3 9.2 54 N/A

a The mix code is an accounting system established to distinguish among the 32 asphalt-aggregate combinations(see Table 3.1).

b Sample ID is specimen or replicate number.

58

Table 4.6. Rut-depths for the OSU Wheel-Tracking Program

Rut-depth, mm a

RC Aggregate

0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00200 2.38 1.54 2.14 2.19 2.22 1.98 1.30 2.08500 4.29 2.51 3.65 3.42 3.19 3.00 2.17 3.15

1,000 6.10 3.89 4.99 4.99 4.52 4.09 2.72 4.472,000 8.06 5.21 6.88 5.59 6.32 5.06 4.48 5.655,000 12.16 7.69 12.29 6.98 8.28 6.65 6.05 7.55

I0,000 24.00 b 10.83 36.00 b 9.87 10.72 9.82 10.17 9.53

RD Aggregate

0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00200 1.03 0.74 1.22 0.77 0.47 0.62 0.39 1.04500 1.72 1.66 2.47 1.66 1.42 1.52 0.92 1.58

1,000 2.22 2.67 3.12 2.54 2.13 2.43 1.32 2.172,000 3.68 3.77 4.35 4.07 3.33 3.99 2.12 3.325,000 5.23 5.68 5.91 5.97 4.96 7.08 3.70 4.56

10,000 6.16 6.84 7.16 7.18 6.31 9.47 4.90 5.19

RH Aggregate

0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00200 1.05 0.63 1.19 0.78 0.80 1.22 0.47 0.95.500 1.86 1.31 1.72 1.42 1.62 2.26 0.93 1.33

1,000 2.88 1.90 2.63 2.26 1.62 3.06 1.05 1.722,000 4.69 3.41 3.71 3.66 3.2 4.22 2.20 2.625,000 6.98 5.87 6.40 5.75 5.58 6.09 3.99 4.41

10,000 I 8.82 7.88 8.68 7.51 7.96 7.70 6.07 6.27

RJ Aggregate

0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00200 0.65 0.49 0.75 0.65 0.60 1.11 0.46 0.59500 1.58 1.04 2.18 1.25 1.40 2.43 1.16 0.95

1,000 2.52 1.99 3.16 1.71 1.77 3.14 1.59 1.282,000 4.42 3.00 4.43 2.49 2.59 4.36 2.48 1.965,000 6.62 3.94 6.91 3.74 4.25 5.81 3.39 2.59

10,000 8.30 4.92 8.79 5.53 6.23 8.65 4.32 2.65

a 1/11. = 25.4mm

b Estimated rut-depth.

59

0'-

-12 I I I0 2,000 4,000 6,000 8,000 10,000

Wheel Passes

AAA-1 AAJ_'IAAC-I AAD-I AAF-1 AA=G-1AA=K-IAAM-ID _ 0 • -= --

Fig. 4.6. OSU wheel-tracking test results for the RC aggregate

-10 •

-12 I I I0 2,000 4,000 6,000 8,000 10,000

Wheel Passes

O 0 _-- -" -- --0

Fig. 4.7. OSU wheel-tracking test results for the RD aggregate

6O

E -4E¢-

c'_ -6

:_ -8fr

-10

-12 I I I I

0 2,00o 4,0o0 6,000 8,000 10,000Wheel Passes

AAA-1 A/_I AAC-1 AAD-1 AAF-1 _ AAK-1 AAM-1[] 0 _: - @

Fig. 4.8. OSU wheel-tracking test results for the RH aggregate0."

-10

-12 I I I I

0 2,000 4,000 6,000 8,000 10,000Wheel Passes

_-ID_,p1_C-lo_,_.1.o_.-1__ _ _,_-1_

Fig. 4.9. OSU wheel-tracking test results for the RJ aggregate

61

oo i _I _ _ i i i i i _ i l i i l l i

Qo M t_- i _i i _ _ i i i i | M_ r : i M l l I l _

l_ i it) _ _-_ _ _ _ _ _ _ _ : _ _c_ _ _ _ i r _ _ _

_ o o

m

62

00_0 O0 OV_ CD 0 O0

_:r.:,_:_ " _.._._.e4°. _. "1,.;,-:

, ,__, _.__, ,__,. _o__

A

_ i_._._ .......... _'_•c__o

• "_& '0 .........

_ m

_ llll

63

experience failure within seven days of testing (approximately 500,000 wheel passes). Voidcontents of the parent beam and test specimen as well as the percent saturation of the testspecimen are also included in Table 4.7. The parent beam is the oversized beam fabricatedat OSU and sent to SWKfUN. SWK/UN subsequently cut the beam to the test specimendimensions. The 10 columns on the right side of the table show the time in hours to attainI, 2, 3, 4, 5, 6, 7, 8, 9, and 10 mm of deformation.

4.4 University of Nevada (Reno) Net Adsorption Test (UNR/NAT)

Program

The NAT results are shown in Table 4.8. The table includes the mean NAT, standarddeviation of the test, and coefficient of variations. The amount of asphalt remaining on the

aggregate indicates how well the aggregate will withstand water-conditioning; lower NATvalues indicate mix that might be water sensitive. Also, the NAT results are showngraphically in Figure 4.10. The NAT shows that aggregate RJ is the most water sensitiveand that aggregate RD is the least water sensitive.

64

Table 4.8. NAT Program

Aggregate Asphalt Mean NAT (%) SD 1 CV2

RC AAA-1 77.05 1.70 2.18AAB-I 76.84 4.00 5.20

AAC-1 80.79 0.20 0.25AAD-1 81.50 0.56 0.70AAF-1 77.80 7.47 9.60AAG-I 78.86 4.32 5.48AAK-I 75.18 2.86 3.80AAM-1 71.90 2.21 3.11

RD AAA-1 74.32 3.30 4.43AAB-1 73.97 2.59 3.50AAC-1 77.63 2.24 2.89AAD-1 81.63 2.49 3.05AAF-I 76.99 3.28 4.27AAG-1 77.17 2.94 3.81AAK-1 81.57 6.66 8.16AAM-I 66.52 3.13 4.17

RH AAA-1 73.29 1.94 2.64AAB-1 74.20 3.65 4.91AAC-1 74.73 2.74 3.66

AAD-I 76.33 1.79 2.34AAF-1 73.06 3.66 5.00AAG-1 55.72 4.86 8.72AAK-1 81.48 3.82 4.69AAM-1 62.23 0.80 1.29

RJ AAA-1 70.09 3.24 4.62AAB-1 63.78 3.31 5.27AAC-1 59.63 3.55 5.96AAD-I 63.50 0.61 0.96AAF-1 56.01 3.60 6.43AAG-I 58.76 8.15 13.87AAK-1 61.57 1.72 2.80AAM-1 58.90 1.45' 2.45

1Standard deviation2Cx_fficient of variation

65

90 t Aggregate

ri Rc tm RD [] RH _ RJ I

o__ 80 ......... I i i ...... : "

_70 - - -

50 ',AAA-1 AAB-1 AAC-1 AAD-1 AAF-I AAG-1 AAK-1 AAM-1

Asphalt Type

Fig. 4.10. NAT results

66

5

Analysis of Results

This chapter presents an analysis of the results summarized in Chapter 4. A description ofthe statistical analyses for the Environmental Conditioning System (ECS), Oregon StateUniversity (OSU) wheel-tracking, SWK Pavement Engineering/University of Nottingham(SWK/UN) wheel-tracking, and University of Nevada (Reno) Net Adsorption Test(NAT/UNR) programs as well as the performance rankings of the materials as determinedby each program are included. A comparison of the performance rankings for eachprogram with those proposed by the A-002A and A-003B contractors, including adiscussion of the results and comparisons is also presented.

5.1 Statistical Analysis

Each test program included 32 asphalt-aggregate mixes according to the experiment designpresented in Chapter 3. The set of 40 tests (32 mixes plus 8 repeated tests) was primarilydesigned to identify the water-sensitivity of the mixes using either rutting (OSU andSWK/UN wheel-tracking) or reduction in modulus (ECS) as the objective function. TheECS test program used full replication (the total of 67 specimens exceeded full replication).The test program provided information to rank the relative performance of the eightasphalts and four aggregates, thus enabling a comparison of results provided by theA-002A, A-003A, and A-003B contractors. The statistical analyses conducted on theresults obtained from the ECS, OSU wheel-tracking, and SWK/UN wheel-trackingprograms are provided in this section.

5.1.1 ECS Test Results

The analysis of the ECS test results employed a general linear model (GLM) procedure toinvestigate the significance of the effect of the different variables and their interactions on

ECS-M R ratio (the dependent variable). GLM procedure uses the method of least squaresto fit general linear models; i.e., test each variable in a given model to see how significantthe variable (or its interaction with other variables) is to the model. Also, GLM can create

67

output data of the dependent variable (ECS-MR) based on the prescribed model; i.e., theoriginal ECS-M R data will be changed to show the effects of the different variables in themodel. One of the statistical methods available in GLM is analysis of variance forunbalanced data that is utilized in ECS analysis. This method was used because the ECStest program has unbalanced data (29 mixes had two replicates and three mixes had threereplicates). GLM procedure is the only statistical method for unbalanced experiments.

The analyses were performed on the results obtained after each of the four conditioningcycles. Table 5.1 shows the variables that were included in the statistical analysis. Thereare two types of independent variables: classification variables (categorical, qualitative,discrete, or nominal variables) and continuous variables (numeric and not necessarilydiscrete). In the model statement any variable that was not defined as a classificationvariable was considered a continuous variable. The aggregate and asphalt type and the time(cycle number) were considered class variables. The other variables were consideredindependent (or covariant) variables.

The analyses were done using an iterative approach. First, a model was selected in whichECS-M R ratio was related to all the variables (Table 5.1) and asphalt-aggregate interactions.Then with each iteration, the least significant variables were removed from the model oneat a time. Table 5.2 shows the results of each iteration; X in front of the variable means

the variable was not significant at 0.05 significance level. The variable that was notsignificant at 0.05 significance level was eliminated from the model in the next iteration.The final model that best represents the effects of asphalt type, initial modulus, andasphalt-aggregate interactions on ECS-M R ratio is shown in Table 5.3.

Table 5.3 shows the output of statistical analysis; the class variables, number of levels, andthe class values are shown. The analysis was performed by time (cycle number); i.e., foreach cycle the model was analyzed (with data for that cycle only). For each cycle, thesummary of the statistical analysis is shown in a separate set of data (Table 5.3).Independent variables (aggregate type, asphalt type, initial modulus, and asphalt-aggregateinteractions) with degree of freedom, type III sum of squares, F-values, and P-values weregiven. For each variable, F-values and P-values (based on type III error) can be checkedfor significance. Type III sum of squares is used to test the significance of each variablebecause the type III test is invariant to the order of variables in the model, and the test ofsignificance for a variable does not involve the parameters of other variables. At time 6initial modulus P-value was 0.0433 and below significance level of 0.05, so initial moduluswas significant to the model at this cycle. For each cycle (time) the model R2, coefficientof variation (CV), and ECS-M R ratio mean are shown. CV gives a relative measure of thevariability in the model in percent; i.e., it can be used to compare one model with another.The given model showed low CV and good R2 values compared with the other models.

68

Table 5.1. Variables Considered in the Analyses of the ECS Test Results

Variable Type Levels

Aggregate type (AGGR) Class RC, RD, RM, RJ

Asphalt type (ASPH) Class AAA-I, AAB-1, AAC=I,AAD-1, AAF-1, AAG-1,AAK-1, and AAM-1

Time (cycle number) Class 6, 12, 18, 24 hr (1, 2, 3, 4cycles)

Percent air-voids (AVOID) Covariant 8 + 1.5

Water-permeability (WK) Covariant 0.0 _ 12.0 E-3 cm/s

Water-permeability ratio (WKR) Covariant 0.03 _ 15.0

Initial air-permeability (AK) Covariant 0.0 _ 20.0 E-5 cm/s

Initial water-permeability (WK0) Covariant 0.0 ,_ 12.0 E-3 cm/s

Initial Modulus Covariant 100 -- 700 ksi

ECS-M R Ratio Dependent 0.6 _ 1.1

69

1. l.

IC I_1 _ _..

_ _ _ -o__o

1 1 1

2

_ z- ± _

_1-_ ___"_ -___ _-____1_ _

°_I_ __I_ : --

e_

_ _ _ "_ = ._

_. _.'- "_ _ _ _ _ "_'_ _ '-'= _

"7O

Table 5.3. GLM Analysis of the ECS Results for Asphalt and Aggregate Type

Class Variables Levels Values

AGGR 4 RC, RD, RH, and RJ

ASPH 8 AAA-1, AAB-1, AAC-1, AAD-I, AAF-1, AAG-1, AAK-1,and AAM-1

Time = 6

Model: R 2 = 0.79, CV = 4.88, ECS-M R Ratio Mean = 0.93

Degree of Type IH Sum Probability ofSource of Error Freedom of Squares F-Values F > Feritie_l

AGGR 3 0.03275601 5.35 0.0037ASPH 7 0.04715846 3.30 0.0079MR0 1 0.00894455 4.38 0.0433AGGR-ASPH 21 0.14340240 3.34 0.0007

Time = 12

Model: R2 = 0.85, CV = 5.22, ECS-M R Ratio Mean = 0.88

Degree of Type EliSum Probability ofSource of Error Freedom of Squares F-Values F > Fcritiesl

AGGR 3 0.07121460 11.13 0.0001ASPH 7 0.04083428 2.73 0.0216MR0 I 0.02653206 12.44 0.0011AGGR-ASPH 21 0.25769088 5.75 0.0001

Time ffi 18Model: R2 = 0.81, CV -- 6.21, ECS-M Ratio Mean = 0.86

Degree of Type Ell Sum Probability ofSource of Error Freedom of Squares F-Values F > Fcritif_l

AGGR 3 0.10603905 12.28 0.0001ASPH 7 0.04310104 2.14 0.0634MR0 1 0.00825944 2.87 0.0987AGGR-ASPH 21 0.23901440 3.95 0.0001

Time = 24Model: R 2 = 0.89, CV = 4.65, ECS-M R Ratio Mean -- 0.84

Degree of Type HI Sum Probability ofSource of Error Freedom of Squares F-Values F > Fcr_,i_,

AGGR 3 0.15659618 33.88 0.0001ASPH 7 0.02909552 2.70 0.0231MR0 1 0.00953970 6.19 0.0175AGGR*ASPH 21 0.23805089 7.36 0.0001

71

However, this analysis does not mean that all the variables that were eliminated did notcontribute to the results of the ECS. The analysis that was done above (Table 5.2) wasperformed for each cycle; i.e., for each cycle the model was tested for variable'ssignificance. In another model in which the analysis was done with cycle number (a classvariable), initial water-permeability was significant to the model.

Another observation was that initial air-permeability was significant to ECS-MR ratio at theend of three cycles. This means that initial fir-permeability influenced the outcome of ECStest results at the end of three cycles. The most important observation from this analysis isthat the asphalt-aggregate interaction is highly significant; i.e., the moisture susceptibility ofone aggregate in a mix is dependent on the type of asphalt and vice versa.

5.1.20SU Wheel-Tracking Test Results

The analysis of the OSU wheel-tracking test results employed a GLM procedure toinvestigate the significance that asphalt type, aggregate type, air-voids, stripping rate, andasphalt-aggregate interaction has on the rut-depth developed after 5,000 wheel passes in theOSU wheel tracker. The results of the analysis are provided in Table 5.4. Unlike theanalysis of ECS test program results, initial analysis of OSU wheel-tracking test resultsshowed that aggregate-asphalt interaction had no effect on rut-depth developed at 5,000wheel passes. The analysis showed very high correlation between rutting at 5,000 wheelpasses and stripping rate, asphalt type, aggregate type, and percent air-voids, at a 0.05significance level (95-percent confidence level).

5.1.3 SWK/UN Wheel-Tracking Test Results

The statistical analysis of the SWK/UN wheel-tracking tests utilized a Bayesian survivalanalysis with time-to-failure distributed as a WeibuU random variable. The Weibull modelemployed a shape factor (C) of 2 (i.e., skewed to the right), a minimum value (A) of zero(A = 0 seemed appropriate since the smallest observed time-to-failure was 2 hours and Amust be less than the smallest observation), and a scale parameter (B) as follows:

_fB e [BAv(i)J AV > 8

= BASPH (J)BAGGR(k);

B = BASPH,(J)BAGGR(k);AV _<8

where:

AV -- percent air-voids of the test specimenBAV(i) = weighting for air-voids with values of 6, 7, 8, 9, or 10

72

Table 5.4. GLM Analysis of the OSU Wheel-Tracking Test Results

Class Levels Values

AGGR 4 RC, RD, RH, and RJ

ASPH 8 AAA-1, AAB-1, AAC-1, AAD-1, AAF-1, AAG-1, AAK-1,and AAM-1

Degree of Type III Sum Probability ofSource of Error Freedom of Squares F-Values F > Fcri_caj

AGGR 3 142.94961295 29.86 0.0001ASPH 7 70.99560815 6.36 0.0001AV2* 1 8.79590144 5.51 0.0234STRIPPING* * 1 10.82167482 6.78 0.0125

• AV2 is air-voids of LCPC-rutted core after OSU wheel-tracking test.• *STRIPPING is visual evaluation of broken specimen after OSU wheel-tracking test.

73

BASPH(j ) = weighting for asphalt type with values of 2, 6, 10, 14, or 18BAGGR(k) = weighting for aggregate type with values of 2, 6, 10, 14, or 18

As shown, the scale parameter is a multiplicative function of asphalt, aggregate, and air-voids with the contribution from air-voids decreasing exponentially for values greater than8 percent and having no contribution (i.e., equal to unity) for air-voids less than or equal to8 percent. It is through the shape parameter (B) that these factors have their effect on thedistribution of time-to-failure.

The SWK/UN wheel-tracking data was tested to determine the probability (Pr) of the time-to-failure (T) being less than or equal to some reasonable time value (in this case sevendays of testing). The test is mathematically represented as follows:

Pr [T<t*] = 1 -e C

where:

A = the minimum allowed time value (zero in this case)B = the scale parameter as previously definedC = the shape factor (2 in this case)t* = predetermined cutoff time value

The above analysis method allows the ranking of asphalt types and aggregate types and atthe same time gives some importance to the air-void content of the test specimen, providedit is greater than 8 percent (i.e., air-void contents greater than 8 percent were considereddetrimental to the probability of the specimen surviving beyond seven days with thedetriment increasing exponentially with air-void content).

The results of the analysis are shown in Table 5.5. For each asphalt and aggregate, thetable lists the probabilities of attaining the scores of 2, 6, 10, 14, and 18 (a range of scoresthat embraces the whole of the data set) and the expected score for the mix components.

The expected score is computed by multiplying the probabilities by their respective scoresand summing the values. Higher expected scores indicate a greater probability of obtaininga pass (i.e., not failing after seven days of testing) in the SWKfUN wheel tracker. Thus, asindicated, the AAM-1 and AAK-1 asphalts and the RC and RD aggregates performed best,and the AAC-1 and AAG-1 asphalts and the ILl aggregate performed worst.

74

Table 5.5. Bayesian Survival Analysis of the SWK/UN Test Results

Mix Probability of Attaining a Score of Expected

Component 2 16 [ 10 ]14 118 Score a

Asphalts

AAA-1 0.0000 0.0225 0.6351 0.2743 0.0681 11.55

AAB-1 0.0000 0.0047 0.3004 0.4293 0.2655 13.82

AAC-1 0.0188 0.9135 0.0606 0.0061 0.0010 6.23

AAD-1 0.0000 0.0000 0.1382 0.4934 0.3683 14.92

AAF-1 0.0000 0.0914 0.5258 0.2806 0.1022 11.57

AAG-1 0.0000 0.7532 0.2252 0.0197 0.0020 7.08

AAK-1 0.0000 0.0000 0.0006 0.1961 0.8032 17.21

AAM-1 0.0000 0.0000 0.0005 0.0143 0.9852 17.94

Aggregates

RC 0.0000 0.0000 0.0948 0.5035 0.4017 15.23

RD 0.0000 0.0000 0.0526 0.6212 0.3262 15.09

RH 0.0000 0.0006 0.4745 0.3930 0.1318 12.62

RJ 0.9862 0.0138 0.0000 0.0000 0.0000 2.06

aExpected score = r_(ProbabilitY)i(score)i; i = 2, 6, 10, 14, 18.

75

5.1.4 NA T/UNR Results

The GLM procedure was used to investigate the effect of aggregate type, asphalt type, andasphalt-aggregate interactions on NAT results. The statistical analysis was one iteration,unlike ECS results analysis (see Table 5.6). Analysis showed that the aggregate-asphaltinteractions were highly significant at 0.05 significance level. Also, the statistical modelused showed high R2 value. Because aggregate-asphalt type interactions are significant,caution must be exercised in interpreting the ranking of aggregate types and asphalt types(similar to ECS).

5.2 Performance Ranking

In addition to investigating which independent variables influence the dependent variablefor each test program, analyses were also performed on the test results with the objective ofranking the materials (asphalts and aggregates) for rutting resistance (OSU and SWK/UNwheel-tracking) and resistance to reduction in modulus (ECS) of moisture-damaged mixes.This section presents the performance rankings of the materials obtained from the analysesof the ECS and the OSU and SWK/UN wheel-tracking test results.

5.2.1 Aggregates

Analysis of the ECS test program results showed the interaction of asphalt type andaggregate type (i.e., ASPH x AGGR) to be significant (Table 5.3). Ranking the ECSresults by aggregate type is inappropriate; thus, aggregate ranking presented in Table 5.7should be interpreted with caution. Ranking of the aggregates based on ECS test resultswas done per cycle, i.e., for each ECS-M R ratio after each cycle. These values are not thearithmetic (true) mean of all ECS-M R ratio values for any given aggregate with the eightasphalts; they are the mean of adjusted ECS-MR ratio values using least squares mean (MRRleast squares mean) (LSMean) for any given aggregate with the eight asphalts. In thestatistical analysis, ECS-M Rratio LSMean was the expected value of ECS-MR ratio if theexperiment was balanced and all the covariant variables were at their mean.

For comparison purposes, it does not make sense to compare one mix with another if themixes have different statistically significant variables values. For example, to compareaggregate RD with aggregate RC, each aggregate specimen has to be adjusted to accountfor the difference in initial modulus (initial modulus is a significant covariant variable) andmust be compared at the same cycle number. The analysis of ECS test results showed that,after three cycles of conditioning, aggregates RH and RD were the best (most moisture-resistant), aggregate RC was in the middle, and aggregate RJ was the worst (most moisture-sensitive). After four conditioning cycles, aggregate RD and RH were still the best, andaggregate RC and RJ were the worst. In Chapter 4 it was mentioned that RC aggregate is

76

Table 5.6. GLM Analysis of the NAT Results for Asphalt and Aggregate Type

Class Variables Levels Values

AGGR 4 RC, RD, RH, and RJ

ASPH 8 AAA-I, AAB-I, AAC-I, AAD-1, AAF-1, AAG-1, AAK-1,and AAM-I

Model: R2 = 0.89, CV = 5.00, NAT mean = 71.57

Degree of Type II1 Sum of Probability of

Source of Error Freedom Squares F-Values F > Ffj.ifie_i

AGGR 3 3725.3464 97.21 0.0001

Table 5.7. Performance Ranking of Aggregates Based on ECS Test

I MRR [MRRAggregate Least Squares Mean Aggregate Least Squares Moan

First Hot Cycle Second Hot Cycle

RD 0.952 RD 0.911RC 0.931 RH 0.897RH 0.921 RC 0.889

RJ 0.899 RJ i 0.840

Third Hot Cycle Freeze Cycle

RH 0.897 RD ] 0.897

RD 0.892 RH [ 0.889

RC 0.860 RC 0.811RJ 0.801 RJ 0.783

77

highly absorptive and tends to disintegrate. The freeze cycle affected aggregate RC (loss instrength) more than all the other aggregates.

Analysis of the OSU wheel-tracking program results showed the interaction of asphalt typeand aggregate type (i.e., ASPH x AGGR) to be not significant. Thus in this case, rankingthe results by aggregate is appropriate. The performance ranking of aggregates (based onleast squares mean) for the OSU wheel-tracking program is summarized in Table 5.8. Asindicated, the analysis showed that the RJ aggregate performed best and the RC aggregateperformed worst. The performance ranking of aggregates based on SWK/UN wheel-tracking test results is summarized in Table 5.9. The ranking indicates that the RC and RDaggregates were good performers and the RJ aggregate was a poor performer. The NATprogram performance ranking of aggregate types in Table 5.10 indicates that aggregates RCand RD performed best (least desorptive) and that aggregate RJ performed worst.

5.2.2 Asphalts

The analysis of results for the ECS test program showed the interaction of asphalt type andaggregate type (i.e., ASPH x AGGR) to be significant. However, ranking the results byasphalt type is inappropriate. The data are shown in Table 5.11. In ranking the asphalttypes, LSMean of ECS-M R ratio was used, similar to the procedure used in aggregateranking. Asphalts AAA-1, AAC-1, and AAB-1 performed better than the other asphalts inthe ECS test; asphalts AAF-1, AAG-1, and AAD-1 showed sensitivity to moisture-damage.

The analysis of results for the OSU wheel-tracking program showed that significance doesnot exist for the asphalt-aggregate interaction. Thus, a ranking by asphalt type can beaccomplished. The performance ranking of asphalts (based on least squares means) for theOSU wheel-tracking program is summarized in Table 5.12. Asphalts AAK-1 and AAM-1performed best (or lowest rut-depth values) and asphalts AAG-1, AAA-1, and AAC-1performed worst (or highest rut-depth values). The performance ranking of asphalts basedon the SWKAJN wheel-tracking test results is summarized Table 5.13. Based on theSWK/LIN wheel-tracking test results, asphalts AAM-1 and AAK-I ranked highest (leastfailures) and asphalt AAC-1 and AAG-1 ranked lowest (most test failures). Theperformance rankings of asphalt types based on NAT results is shown in Table 5.14.Asphalt AAD-1 ranks best (lowest desorption values), and asphalts AAG-1 and AAM-1ranks worst (highest desorption values).

5. 2. 3 Mixes

The statistical analysis of the ECS results shows that the asphalt-aggregate interaction isvery significant based on 0.05 significance level (i.e., 95-percent confidence). Thisconclusion rejects any rankings by asphalt type only or aggregate type only. To concludethat aggregate RD performs much better than RJ in moisture susceptibility, a singlecommon asphalt would have to be matched with either of these aggregates. The statisticalanalysis of OSU wheel-tracker results showed no asphalt-aggregate interactions, so it wouldbe inappropriate to include rankings based on mixes here. Table 5.15 shows ECS rankingbased on ECS-M Rratio after each cycle, and the mixes are ranked from 1 to 32.

78

Table 5.8. Performance Ranking of Aggregates (OSU Wheel-Tracking Program)

Level Least Squares Mean Homogenous Groups Pexformance Ranking

RJ 4.456875 A Good

RD 5.384375 B IntermediateRI-I 5.653125 B

RC 8.475375 C Poor

Table 5.9. Performance Ranking of Aggregates (SWK/UN Wheel-Tracking Program)

Level Least Squares Mean Homogenous Groups Performance Ranking

RC 15.23 A GoodRD 15.09 A

RH 12.62 B Intermediate

1LI 2.06 C Poor

Table 5.10. Performance Ranking of Aggregates (NAT/UNR Test Program)

Level Least Squares Mean Homogenous Groups Performance Ranking

RC 77.49 A GoodRD 76.05 A

RH 71.39 B Intermediate

ILl 61.53 C Poor

79

Table 5.11. Performance Ranking of Asphalt Based on ECS TestI

MRR Least [ Least Squares MRR Least Least Squares

I

Asphalt Squares Mean [ Mean Number Asphalt Squares Mean Mean Number

Asphalt First Hot C rde Second Hot Cycle

AAB-1 0.968 1 AAC-I 0.924 1AAA-1 0.956 2 AAA-1 0.922 2AAC-1 0.934 4 AAB-1 0.895 3AAF-1 0.926 5 AAG-1 0.874 4AAG-1 0.923 5 AAM-1 0.867 5AAD-1 0.910 6 AAF-1 0.865 6AAM-I 0.910 7 AAK-1 0.865 7AAK-1 0.880 8 AAD-1 0.861 8

Third Hot Cycle Freeze Cycle

AAA-1 0.921 1 AAA-I 0.901 1AAB- 1 0.894 2 AAC-1 0.868 2AAC-1 0.894 3 AAB-1 0.851 3AAM-1 0.855 4 AAK-1 0.846 4AAK-1 0.840 5 AAM-1 0.831 5AAD-1 0.834 6 AAG-1 0.825 6AAF-1 0.834 7 AAD-1 0.816 7AAG-1 0.828 8 AAF-1 0.809 8

I II

Table 5.12. Performance Ranking of Asphalts (OSU Wheel-Tracking Program)

Level Least Squares Mean Homogenous Groups a Performance Ranking

AAK-1 4.28125 A GoodAAM-I 4.77750 AB Good

AAD- 1 5.60875 BC Intermediate

AAF-1 5.76625 BC IntermediateAAB-1 5.79375 BC Intermediate

AAG-1 6.40500 C Poor

AAA-1 7.38625 C PoorAAC-1 7.87750 C Poor

a Groups with the same letter designation are not significantly different.

80

Table 5.13. Performance Ranking of Asphalts (SWK/UN Wheel-Tracking Program)

Level Expected Score Homogenous Groups Performance Ranking

AAM-1 17.94 A Very GoodAAK-1 17.21 A

AAD-1 14.92 B GoodAAB-1 13.82 B

AAF-1 11.57 C FairAAA-1 11.55 C

AAG-1 7.08 D PoorAAC-1 6.23 D

Table 5.14. Performance Ranking of Asphalts (NAT/UNR Test Program)

Level Least Squares Mean Homogenous Groups Performance Ranking

AAD-1 75.487 A Good

AAK-1 74.950 A B IntermediateAAA-1 73.688 A BAAC-1 73.198 A BAAB-1 72.199 A BAAF-1 70.964 B

AAG-1 66.759 C PoorAAM-1 64.912 C

81

Table 5.15. Ranking of 32 Mixes after Each ECS Cycle

ECS MR Least ECS M R LeastLeast Squares Least SquaresSquares Mean Squares Mean

Aggregate Asphalt Mean Number Aggregate Asphalt Mean Numberi i

First Hot Cycle Second Hot Cycle

RH AAC-1 1.090 1 RH AAC-1 1.170 1RJ AAF-1 0.993 2 RJ AAF-1 0.957 2RH AAB-1 0.985 3 RE) AAA-1 0.953 3RD AAA-1 0.980 4 RH AAD-1 0.950 4RD AAD-1 0.975 5 RC AAA-1 0.945 5RC AAG-1 0.975 6 RD AAC-1 0.940 6RC AAB-1 0.970 7 RC AAG-1 0.935 7RC AAA-1 0.970 8 RD AAK-1 0.935 8RD AAC-1 0.965 9 RC AAM-1 0.920 9R/ AAB-1 0.965 10 RD AAM-1 0.915 10RC AAF-1 0.965 11 RC AAB-1 0.905 11RC AAM-1 0.960 12 RH AAB-1 0.905 12RD AAM-I 0.960 13 RE) AAB-1 0.903 13RH AAD-1 0.955 14 RH AAA-1 0.900 14RD AAK-1 0.950 15 RD AAG-1 0.897 15RD AAB-I 0.950 16 RJ AAA-1 0.890 16RH AAA-I 0.940 17 RH AAG-1 0.890 17RJ AAA-I 0.935 18 RD AAD-1 0.885 18RD AAG-1 0.930 19 RC AAK-1 0.885 19RJ AAM-1 0.915 20 RJ AAM-1 0.875 20RD AAF-1 0.907 21 RC AAF-1 0.870 21RJ AAG-I 0.905 22 RJ AAB-1 0.865 22RC AAK- 1 0.895 23 RD AAF- 1 0.857 23RJ AAG-I 0.880 24 RC AAC-1 0.840 24RC AAC-1 0.865 25 RH AAK-1 0.830 25

P.J AAD-1 0.860 26 RC AAD-1 0.810 26RC AAD-1 0.850 27 KI AAK-1 0.810 27RH AAK-1 0.845 28 RJ AAD-1 0.800 28RH AAF-I 0.840 29 RH AAF-I 0.775 29RJ AAK-I 0.830 30 RJ AAG-I 0.775 30

RJ AAC-I 0.815 31 RH AAM-1 0.757 31RH AAM-1 0.807 32 RJ AAC-1 0.745 32

Continued on next page.

82

Table 5.15. Ranking of 32 Mixes after Each ECS Cycle (continued)

ECS MR Least ECS Ms LeastLeast Squares Least SquaresSquares Mean Squares Mean

Aggregate Asphalt Mean Number Aggregate Asphalt Mean Number

Third Hot Cycle Freeze Cycle

RH AAC- 1 1.125 1 RH AAC- 1 I. 125 1RH AAA-1 0.950 2 RD AAK-1 0.955 2RD AAA- 1 0.943 3 RH AAA- 1 0.940 3RD AAK-1 0.930 4 RD AAA-1 0.933 4RH AAB-1 0.925 5 RD AAG-1 0.927 5RC AAM-1 0.920 6 RH AAB-1 0.910 6RH AAD-1 0.915 7 RH AAD-1 0.910 7RD AAB-1 0.907 8 RD AAM-1 0.890 8RC AAA-1 0.905 9 RD AAC-1 0.885 9RD AAC-1 0.905 10 RC AAM-1 0.885 10RJ AAF-1 0.903 11 RH AAG-1 0.880 11RC AAB-1 0.895 12 RJ AAA-1 0.870 12RD AAM-1 0.895 13 RC AAA-1 0.860 13RC AAG-1 0.885 14 RD AAB-1 0.860 14RJ AAA-1 0.885 15 RD AAD-1 0.845 15RH AAG-1 0.885 16 RC AAK-I 0.840 16RD AAG-I 0.873 17 RJ AAF-1 0.840 17RC AAK-1 0.870 18 RD AAF-1 0.830 18RJ AAB- 1 0.850 19 RJ AAB- 1 0.820 19RD AAD- 1 0.845 20 RC AAB- 1 0.815 20RD AAF-1 0.837 21 RC AAG-I 0.810 21RC AAC- 1 0.830 22 RJ AAK- 1 0.805 22RJ AAM-1 0.825 23 RH AAF-1 0.795 23RH AAF- 1 0.800 24 RH AAK- 1 0.785 24RH AAK- 1 0.795 25 RJ AAM- 1 0.785 25RC AAF- 1 0.795 26 RC AAF- 1 0.770 26RJ AAD- 1 0.795 27 RH AAM- 1 0.763 27RH AAM- 1 0.780 28 RC AAD- 1 0.760 28

RC AAD- 1 0.780 29 RJ AAD- 1 0.750 29RJ AAK-1 0.765 30 RC AAC-1 0.750 30RJ AAC-1 0.715 31 RJ AAC-1 0.710 31RJ AAG-1 0.670 32 ILI AAG-1 0.685 32

83

The data presented in Table 5.15 are based on LSMean procedure of GLM statisticalmethod, similar to ranking of asphalts and aggregate types. Table 5.15 does not show thebreakdown between poor aggregate (moisture-susceptible) and good aggregate (moisture-resistant) or between poor asphalt and good asphalt. The mixes are not grouped byhomogenous groups, in which mixes within the same group are not significantly differentand each group is ranked. However, it shows the breakdown between moisture-susceptiblemixes and moisture-resistant mixes. After each cycle, mixes that tended to be moisture-susceptible progressively lost stiffness, but the mixes that were least susceptible tomoisture-damage maintained about the same stiffness. Table 5.15 shows that some mixesthat performed well after one cycle did not maintain the same rmlking after subsequentcycles. Figure 5.1 shows the ranking of the 32 mixes (based on LSMean of ECS-M Rratio)after one and three conditioning cycles. The significance of this observation is that onecycle of ECS conditioning is not sufficient and results are unpredictable (ranking of themixes might not have a good basis). Finally, note that the range of data presented in Table5.15 is very small; i.e., ECS-MR ratio of all 32 mixes varies between 1.12 and 0.685.

5.3 Comparison of A-002A and A-003B Results

This section compares A-003A results with A-002A and A-003B (including UNR/NAT testprogram) results pertaining to moisture sensitivity and permanent deformation. The rankingprovided by A-003B covered four aggregates and three asphalts. Since the rankings ofNATs performed by both A-003B and UNR were similar and since A-003B had only threeasphalts, both NAT rankings were combined. The rankings of aggregates from A-002A,A-003A, and A-003B contractors are summarized in Table 5.16. It is evident that theSWK/UN wheel-tracking results and the A-003B UNR/NAT results are in agreement.However, disparity exists between the OSU wheel-tracking results and those of the othertwo tests.

The rankings of asphalts from the A-002A, A-003A, and A-003B (including UNR/NAT)contractors are summarized in Table 5.17. As indicated, the SWK/UN wheel-tracking

results, the OSU wheel-tracking results, and the A-002A predictions for permanentdeformation are in agreement. Thus, it appears that the OSU and SWK/UN wheel-trackingtest results do indeed validate the predictions made by the SEC-tan delta tests proposed by

A-002A. Figure 5.2 shows the NAT and ECS-M R results plotted against the ECS rankings.The plot shows that about 16 of the 32 mixes have similar ranking and that the rest aredifferent.

However, very little agreement exists between the wheel-tracking test results and thepredictions for the water-sensitivity of the binder made by the A-002A and A-003Bcontractors and, since both wheel-tracking tests indicate similar rankings that closely matchthe predicted ranking for permanent deformation, it appears that either the wheel-trackingtests are poor indicators of water-sensitivity of the binder or that the NAT (A-003B) andthe FTIR test (A-002A) are not appropriate tests for predicting the water-sensitivity of thebinder. It is appropriate to point out that the original performance predictions for water-

84

1.2

-_ One cycle _- Three cycles

O

l.0

&_0.8

0.6 i , , , , , , , , , , , , , , r

Asphalt-Aggregate Mixture Ranking

Fig. 5.1. Ranking of 32 mixes after one and three cycles

12 90

• NAT_,m 2_.,S1.1

O

:,= o-,9.1.0n,"-75 ="O

"' 0.9 .......... '-"_ OO

"a, <

_0.8 ............... ._ouJ 60 z

1 o.7 ................................. 55

0.6 50

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31Asphalt-Aggregate Mixture Ranking

Fig. 5.2. Ranking of 32 mixes by NAT and ECS test

85

Table 5.16. Snmmary of Aggregate Rankings

Water-sensitivity Rutting

Performance A-003B/UNR A-003A A-003ARanking (Net adsorption) (OSU wheel-tracking) (SWK/UN wheel-tracking)

Good RC, RD RJ RC, RDRH RD, RH RH

Poor RJ RC RJ

Table 5.17. Summary of Asphalt Rankings

Water-sensitivity Permanent Deformation

A-003A A-003APerformance (OSU wheel- (SWK/UN wheel-Ranking A-002A A-003B/UNR A-002A tracking) tracking)

Good AAF-1 AAD-1 AAM-1 AAK-1 AAM-1AAB-1 AAK-1 AAK-1 AAM-1 AAK-1AAM-1AAA-1 AAA-1 AAD-1 AAD-1 AAD-1

AAC-1 AAB-1 AAF-1 AAB-1AAD-1 AAB-1 AAB-1AAK-1 AAF-1 AAA-1 AAF-1

AAF-1 AAG-1 AAA-1AAG-1 AAG-1 AAC-1 AAA-1

AAM-1 AAG-1 AAC-1 AAG-1Poor AAC-1

I

86

sensitivity proposed by the A-002A contractor were tentative and were later withdrawn;therefore, there is no hypothesis for water-sensitivity prediction by A-002A.

Considering the comparisons of ranking of the materials by different test procedures, keepin mind that the mechanisms leading to varying performance are not the same. The aim oftesting reported herein was to measure water-sensitivity, but all the tests did not do sodirectly. Both the ECS test and NAT evaluated the mix before and after water-conditioning, but the rutting tests of OSU and SWK/UN evaluated the mix only after thewet conditioning. Because of the large specimen size of the beams tested (compared withECS or NAT specimens), the water-conditioning of the beams may not have been severeenough to induce true water damage. Furthermore, the NAT procedure addressed only thepotential for stripping (adhesion) and was not capable of evaluating cohesion loss. Theother tests (ECS and OSU and SWKAJN wheel-tracking) included all the mechanismssimultaneously and provided a gross effect without clearly separating the cause of failure ineach case.

5.4 Discussion Regarding Specifications

One of the goals of the ECS test development was to establish guidelines for specifications.The appropriate limits for specifications should be based on field performance. Thevalidation testing covered in this report includes only Materials Reference Library (MRL)materials with no direct link to field projects. Therefore, specifications are discussed in thereport on field validation (Allen and Terrel 1994), which includes actual pavements, andthus provides an opportunity to compare field and laboratory results.

87

6

Conclusions and Recommendations

6.1 Conclusions

The testing results and analysis presented herein warrant the following conclusions:

1) Performance ranking of mixes by asphalt type or aggregate type alone cannotbe made for the Environmental Conditioning System (ECS) test results becauseof the significant interaction between asphalt and aggregate. Water-sensitivityin the ECS is significant for pairs of asphalts and aggregate. The ECS testresults show that ECS performance ranking after only one cycle is notsufficient and does not correlate with ranking after three cycles.

2) The Oregon State University (OSU) wheel-tracking test results indicate that theRJ aggregate is a good performer, the RC aggregate a poor performer, and theRD and RH aggregates intermediate performers in terms of rut resistance.The SWK Pavement Engineering/University of Nottingham (SWK/UN) wheel-tracking test results indicate that the RC and RD aggregates are goodperformers (with practically no difference between the two), the RH aggregatean intermediate performer, and the RJ aggregate a poor performer. Thesignificant differences between the results of the two test methods maypossibly be attributed to the significant differences in testing methods, testapparatus, specimen size, and specimen environment during testing. However,the results of the SWK/UN wheel-tracking test generally validate thepredictions proposed by the net adsorption test (NAT) (A-003B), but those ofthe OSU wheel-tracking test do not. Thus, it appears that the OSU wheel-tracking test may not be appropriate for evaluating aggregate type as it pertainsto water-sensitivity.

3) The SEC-tan delta test proposed by A-002A adequately predicts the ruttingpotential of asphalt by type as evidenced by close agreement with the asphaltrankings from the OSU wheel-tracking and SWK/UN wheel-tracking tests.Near perfect agreement exists between A-002A predictions and the SWK/UNresults unless both dry and wet-conditioned specimens are included.

89

4) Predictions of the water-sensitivity of the binder as proposed by the A-002Ab'TIR test and the A-003B NAT show tittle or no correlation to wheel-trackingtests on the mixes. Since very good to excellent correlation exists among thewheel-tracking tests and the A-002A predictions for permanent deformation, itappears that the FTIR test and the NAT are lxx_r indicators of the moisturesensitivity of the binder.

6.2 Recommendations

It was evident from the results of this research that some of the test procedures used are notappropriate for evaluating water-sensitivity of mixes. Therefore, recommendations forimproving comparisons in future research follow:

1) The ECS should be used to evaluate only specific pairs (i.e., asphalt-aggregatecombinations), using at least three conditioning cycles.

2) If water-sensitivity is important in the OSU wheel-tracking tests, both dry andwater-conditioned specimens should be tested. This approach will provide aratio of wet to dry rutting (and possibly other failures) similar to that of theECS.

3) An improved method of water-conditioning must be developed for the largebeam specimens used in the OSU wheel tracker. The method used in thisproject was slow and cumbersome, and the thoroughness of wetting and/orconditioning was uncertain.

90

7

References

A1-Joaib, A. (1993). Evaluation of Water Damage on Asphalt Concrete Mixes Using theEnvironmental Conditioning System. Transportation Research Record,Transportation Research Institute, Oregon State University, May 1993.

Allen, W., and R. L. Terrel (1994). Field Validation of the Environmental ConditioningSystem. Report no. SHRP-A-396. Strategic Highway Research Program, NationalResearch Council, Washington, D.C.

Curtis, C. W., K. Ensley, and J. Epps (1993). Fundamental Properties ofAsphalt-Aggregate Interactions Including Adhesion and Adsorption. Report No.SHRP-A-341. Strategic Highway Research Program, National Research Council,Washington, D.C.

Curtis, C. W., L. M. Perry, and C. J. Brannan (1991). An Investigation ofAsphalt-Aggregate Interactions and Their Sensitivity to Water. Proceedings,Strategic Highway Research Program and Traffic Safety on Two Continents,Gothenburg, Sweden.

Robertson, R. E. (1991). Letter to James Moulthrop, A-001 Technical Program Director,regarding updated rankings of SHRP asphalts by chemical methods.

SAS Institute Inc. (1988). SAS/STAT Users Guide, Release 6.03, Cary, N.C.

Terrel, R. L., and S. A1-Swailmi (1994). Water-Sensitivity of Asphalt-Aggregate MixturesTest Development. Final Report, SHRP A-003B. Strategic Highway ResearchProgram, National Research Council, Washington, D.C.

Terrel, R. L., and J. W. Shute (1989). Summary Report on Water-Sensitivity. Report no.SHRP-A-304. Strategic Highway Research Program, National Research Council,Washington, D.C.

91

Appendix A

Test Data

93

. . o___--_ _o_--___ _

_N4_g__d_ dddd44Mddd4_Nd

N

" 4_d44_4444Nd44 NNNNdNNNNNN_N

0,o°0°°.°°°°o°o °o°°0,°°0o°° °

ooo_ oooo_ooooo oo_o ooo o.... . " " "_*** _o* _*

94

o.°oooooo**.**o ooo°°°ooooo°°

.......... _, .... t'_l ''_' t._.. C_,lt_. _ .............

•-+ 0.1 ,-_ 0,1 ,._ ,'_ v.'.'+,-+ ,..'.+ ('_1 ,-+ 0 0 ,-+ 0 ,-+ ,.'.'_ ,.-'_ 0 _'_1 ,-+ ,-+ ,-_ ,-_ 0,1 ,-+ _-+ 0

****°.°+°oo°oo° _

95

ECS DataII

Air Cond ECS ECS Water Retained

Asph Agg Voids Time M r Mr Perm PermCode Code (9_) (hr) (ksi) Ratio (E"3 cm/s) Ratio

AAF RJ 8.4 0 473.0 1.00 Very Low N/A6 470.2 0.99 Very Low N/A

12 468.0 0.99 Very Low N/A18 453.1 0.96 Very Low N/A24 403.6 0.85 Very Low N/A

AAC RJ 6.4 0 220.0 1.00 3.81 1.006 189.0 0.86 2.91 0.76

12 174.0 0.79 1.08 0.2818 164.5 0.75 0.13 0.0424 164.0 0.75 0.10 0.03

AAM RJ 8.2 0 318.0 1.00 2.13 1.006 278.7 0.88 1.98 0.93

12 262.8 0.83 1.72 0.8118 251.7 0.79 0.96 0.4524 242.1 0.76 0.53 0.25

AAK RJ 8.7 0 255.0 1.00 3.65 1.006 218.2 0.86 3.48 0.95

12 213.4 0.84 3.54 0.9718 204.7 0.80 3.30 0.9024 212.6 0.83 3.30 0.90

AAK RJ 8.2 0 275.0 1.00 4.87 1.006 219.0 0.80 3.93 0.81

12 215.0 0.78 3.14 0.6418 201.5 0.73 3.07 0.6324 213.4 0.78 3.44 0.71

AAB RJ 8.2 0 210.0 1.00 5.04 1.006 197.5 0.94 1.91 0.38

12 190.3 0.91 0.93 0.1818 189.5 0.90 0.10 0.0224 177.6 0.85 0.10 0.02

AAD RJ 7.5 0 215.0 1.00 3.39 1.00

6 174.4 0.81 2.75 0.8112 172.0 0.80 0.12 0.0418 160.9 0.75 0.07 0.0224 157.7 0.73 0.05 0.02

AAA RJ 8.3 0 155.0 1.00 2.28 1.006 137.8 0.89 2.42 1.06

12 136.2 0.88 1.77 0.7818 135.0 0.87 0.59 0.2624 133.0 0.86 0.11 0.05

96

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr Mr Perm PermCode Code (%) (hr) 0csi) Ratio (E-3 cm/s) Ratio

AAF RJ 8.1 0 550.0 1.00 0.08 1.00

6 547.5 1.00 Very Low N/A12 502.5 0.91 Very Low N/A18 473.9 0.86 Very Low N/A24 438.8 0.80 Very Low N/A

AAG RJ 9.4 0 440.0 1.00 7.72 1.006 350.4 0.80 4.73 0.61

12 298.7 0.68 4.58 0.5918 297.9 0.68 4.10 0.5324 297.1 0.68 3.89 0.50

AAA RJ 8.1 0 136.0 1.00 1.89 1.006 133.0 0.98 0.10 0.05

12 122.6 0.90 0.10 0.0518 122.0 0.90 0.09 0.0524 120.3 0.88 0.04 0.02

AAA RD 8.0 0 195.0 1.00 2.20 1.006 190.7 0.98 4.36 1.98

12 190.7 0.97 3.19 1.4518 185.0 0.95 2.92 1.3324 180.5 0.93 2.90 1.32

AAC RD 8.6 0 245.0 1.00 10.53 1.006 240.0 0.98 7.34 0.70

12 227.0 0.93 7.01 0.6718 215.4 0.88 6.92 0.6624 214.6 0.88 6.92 0.66

AAD RD 9.2 0 218.0 1.00 8.57 1.006 216.0 0.99 5.46 0.64

12 192.0 0.88 3.19 0.3718 178.8 0.82 4.36 0.5124 178.8 0.82 4.26 0.50

AAA RD 8.1 0 177.0 1.00 2.89 1.006 175.0 0.99 3.36 1.16

12 165.2 0.93 3.27 1.1318 164.3 0.93 3.27 1.1324 165.2 0.93 3.27 1.13

AAD RD 8.8 0 195.0 1.00 5.82 1.006 187.0 0.96 5.36 0.92

12 173.7 0.89 5.20 0.8918 170.0 0.87 5.20 0.8924 170.3 0.87 5.20 0.89

97

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr M r Perm PermCode Code (%) Oar) (ksi) Ratio (E-3 cm/s) Ratio

AAG RD 8.0 0 510.0 1.00 3.24 1.006 440.0 0.86 2.32 0.72

12 430.0 0.84 1.60 0.4918 408.7 0.80 1.55 0.4824 466.8 0.92 1.50 0.46

AAB RD 7.2 0 280.0 1.00 4.08 1.006 259.0 0.93 3.85 0.94

12 234.2 0.84 2.83 0.6918 215.4 0.77 2.83 0.69

24 214.0 0.76 2.03 0.50

AAG RD 7.7 0 540.0 1.00 0.05 1.006 530.0 0.98 0.19 3.73

12 490.5 0.91 0.09 1.8218 481.1 0.89 0.09 1.8224 502.4 0.93 0.09 1.82

AAF RD 9.6 0 560.0 1.00 4.99 1.006 545.0 0.97 5.95 1.19

12 489.6 0.87 5.76 1.1518 457.7 0.82 5.33 1.0724 450.0 0.80 5.28 1.06

AAG RJ 8.1 0 265.0 1.00 3.97 1.006 254.8 0.96 0.72 0.18

12 231.0 0.87 0.12 0.0318 175.2 0.66 0.07 0.0224 184.0 0.69 0.07 0.02

AAG RH 6.8 0 640.0 1.00 0.05 1.006 553.5 0.86 2.52 48.46

12 545.0 0.85 0.13 2.4018 540.7 0.84 0.09 1.6924 553.5 0.86 0.05 0.88

AAK RD 8.7 0 293.0 1.00 2.71 1.006 274.2 0.94 3.48 1.28

12 269.1 0.92 3.58 1.3218 281.0 0.96 3.73 1.3824 280.0 0.96 3.73 1.38

AAA RH 7.5 0 135.0 1.00 6.03 1.006 128.0 0.95 4.41 0.73

12 125.0 0.93 4.41 0.7318 130.0 0.96 2.75 0.4624 128.0 0.95 3.40 0.56

98

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr Mr Perm PermCode Code (%) (hr) (ksi) Ratio (rE-3 cm/s) Ratio

AAK RD 8.1 0 287.0 1.00 2.13 1.006 275.0 0.96 3.32 1.56

12 273.0 0.95 3.32 1.5618 258.9 0.90 3.12 1.4624 272.5 0.95 3.12 1.46

AAG RH 5.9 0 610.0 1.00 0.05 1.006 • 580.0 0.95 2.13 42.60

12 566.0 0.93 0.13 2.5418 566.0 0.93 0.09 1.7824 549.2 0.90 0.08 1.68

AAF RH 7.6 0 454.0 1.00 0.16 1.00

6 388.3 0.86 2.69 17.1312 387.5 0.85 2.26 14.3918 383.3 0.84 2.12 13.5024 383.0 0.84 2.12 13.50

AAM RH 7.1 0 430.0 1.00 0.00 N/A6 365.0 0.85 4.37 1.00

12 344.6 0.80 0.12 0.0318 374.9 0.87 2.83 0.6524 368.3 0.86 2.73 0.62

AAA RH 8.4 0 118.0 1.00 5.66 1.006 110.3 0.93 4.83 0.85

12 102.3 0.87 4.16 0.7308 110.5 0.94 4.16 0.7324 109.4 0.93 4.16 0.73

AAB 1Ll 8.5 0 465.0 1.00 4.03 1.006 460.0 0.99 1.41 0.35

12 382.0 0.82 0.15 0.04

18 373.8 0.80 0.17 0.0424 368.6 0.79 O.15 0.04

AAF ILl 9.1 0 256.0 1.00 5.54 1.006 254.2 0.99 2.64 0.48

12 248.6 0.97 2.13 0.3818 228.7 0.89 0.93 0.1724 222.7 0.87 0.13 0.02

AAB RD 6.8 0 300.0 1.00 0.00 N/A6 283.5 0.95 3.92 1.00

12 281.8 0.94 2.58 0.6618 293.8 0.98 1.98 0.5124 268.2 0.89 0.86 0.22

99

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr Mr Penn PennCode Code (%) (hr) (ksi) Ratio (E-3 cm/s) Ratio

AAM RD 10.1 0 285.0 1.00 2.70 1.006 283.5 0.99 3.35 1.24

12 269.4 0.95 2.79 1.0318 269.0 0.94 3.16 1.1724 247.0 0.87 3.16 1.17

AAA RD 8.2 0 190.0 1.00 0.67 1.006 184.3 0.97 2.48 3.70

12 182.0 0.96 2.48 3.7018 180.0 0.95 2.20 3.28

24 178.8 0.94 2.00 2.99

AAM RC 9.7 0 235.0 1.00 13.18 1.006 223.2 0.95 8.58 0.65

12 210.0 0.89 7.09 0.5418 210.0 0.89 5.94 0.4524 204.2 0.87 5.76 0.44

AAK RC 9.4 0 250.0 1.00 8.93 1.006 215.5 0.86 5.12 0.57

12 216.0 0.86 4.57 0.51

18 212.0 0.85 4.48 0.5024 209.0 0.84 4.22 0.47

AAM RJ 9.0 0 280.0 1.00 2.73 1.006 266.9 0.95 2.30 0.84

12 258.6 0.92 2.30 O.8418 240.0 0.86 2.24 0.8224 226.0 0.81 1.20 0.44

AAD RH 7.6 0 172.0 1.00 0.00 N/A6 162.0 0.94 0.17 1.00

12 161.0 0.94 0.17 0.9918 152.0 0.88 0.16 0.8924 150.0 0.87 0.14 0.81

AAK RH 8.4 0 248.0 1.00 3.27 1.006 210.0 0.85 3.56 1.09

12 208.0 0.84 3.39 1.0418 202.0 0.81 2.69 0.8224 198.0 0.80 2.38 0.73

AAB RH 7.4 0 250.0 1.00 0.11 1.006 250.0 1.00 2.11 19.01

12 232.0 0.93 2.11 19.0118 232.0 0.93 2.11 19.0124 222.0 0.89 1.77 15.95

100

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr Mr Penn PermCode Code (%) Car) 0csi) Ratio (E"3 cm/s) Ratio

AAF RH 6.9 0 675.0 1.00 0.00 N/A

6 555.0 0.82 0.13 1.00

12 475.0 0.70 0.17 1.3618 510.0 0.76 0.19 1.4924 505.0 0.75 0.16 1.29

AAC RD 8.6 0 285.0 1.00 9.33 1.00

6 270.0 0.95 7.09 0.7612 270.0 0.95 6.48 0.6918 265.0 0.93 5.96 0.6424 255.0 0.89 5.96 0.64

AAC RH 7.0 0 230.0 1.00 0.00 N/A6 240.0 1.04 0.13 1.00

12 275.0 1.20 0.11 0.8318 260.0 1.13 0.06 0.4924 260.0 1.13 0.05 0.42

AAM RH 6.8 0 400.0 1.00 0.00 N/A6 327.0 0.82 0.20 1.00

12 300,0 0.75 0.16 0.80

18 299.0 0.75 0.16 0.80

24 286.0 0.72 0.15 0.77

AAA RC 8.3 0 220.0 1.00 3.55 1.006 210.0 0.95 2.72 0.77

12 210.0 0.95 2.26 0.64

18 199.0 0.90 2.22 0.63

24 183.0 0.83 2.22 0.63

AAB RC 9.2 0 255.0 1.00 5.42 1.006 245.0 0.96 4.09 0.75

12 242.0 0.95 3.44 0.6318 239.0 0.94 3.44 0.6324 215.0 0.84 3.14 0.58

AAD RC 9.2 0 230.0 1.00 3.72 1.006 195.0 0.85 3.91 1.05

12 179.0 0.78 3.60 0.9718 178.0 0.77 3.29 0.8824 174.0 0.76 3.15 0.85

AAD RC 8.7 0 246.0 1.00 0.03 1.006 209.0 0.85 0.15 4.93

12 206.0 0.84 0.13 4.2318 194.0 0.79 0.12 3.9324 188.0 0.76 0.12 3.83

101

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr Mr Perm PermCode Code (%) (hr) (ksi) Ratio (E-3 cm/s) Ratio

AAC RC 9.0 0 335.0 1.00 4.92 1.006 275.0 0.82 4.20 0.85

12 270.0 0.81 3.63 0.7418 270.0 0.81 3.20 0.65

24 250.0 0.75 2.67 0.54

AAC RC 9.0 0 275.0 1.00 5.00 1.006 250.0 0.91 3.17 0.63

12 240.0 0.87 2.77 0.5518 233.0 0.85 2.21 0.44

24 207.0 0.75 1.89 0.38

AAK RC 9.2 0 280.0 1.00 5.89 1.006 260.5 0.93 4.23 0.72

12 255.5 0.91 3.47 0.5918 250.0 0.89 2.80 0.4824 235.0 0.84 2.55 0.43

AAF RC 8.3 0 490.0 1.00 2.24 1.006 470.0 0.96 0.70 0.31

12 458.0 0.93 0.11 0.0518 385.0 0.79 0.08 0.0424 374.0 0.76 0.04 0.02

AAG RC 10.1 0 410.0 1.00 10.31 1.006 398.0 0.97 5.42 0.53

12 378.0 0.92 4.36 0.4218 373.0 0.91 3.68 0.3624 326.0 0.80 2.31 0.22

AAG RC 10.4 0 315.0 1.00 7.63 1.006 310.0 0.98 4.56 0.60

12 299.0 0.95 3.87 0.5118 270.0 0.86 3.25 0.4324 258.0 0.82 2.22 0.29

AAF RC 9.0 0 481.0 1.00 9.42 1.006 466.0 0.97 4.35 0.46

12 388.0 0.81 4.16 0.4418 385.0 0.80 3.54 0.3824 375.0 0.78 3.21 0.34

AAM RC 10.5 0 275.0 1.00 6.02 1.006 267.0 0.97 3.24 0.54

12 262.0 0.95 2.73 0.4518 261.0 0.95 2.41 0.4024 248.0 0.90 2.27 0.38

102

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time Mr Mr Perm PermCode Code (%) (hr) (ksi) Ratio (E"3 cm/s) Ratio

AAD RJ 7.5 0 155.0 1.00 4.08 1.006 141.0 0.91 0.97 0.24

12 124.0 0.80 0.14 0.0318 130.0 0.84 0.14 0.0324 120.0 0.77 0.10 0.02

AAB RC 9.5 0 250.0 1.00 3.93 1.006 246.0 0.98 2.97 0.76

12 214.0 0.86 2.11 0.5418 213.0 0.85 2.07 0.53

24 198.0 0.79 1.78 0.45

AAA RC 9.0 0 160.0 1.00 5.27 1.006 158.0 0.99 4.44 0.84

12 150.0 0.94 3.52 0.6718 146.0 0.91 3.52 0.6724 142.0 0.89 2.89 0.55

AAB RH 9.1 0 210.0 1.00 0.00 N/A6 203.0 0.97 2.89 1.00

12 185.0 0.88 2.07 0.7218 193.0 0.92 2.07 0.7224 195.0 0.93 1.81 0.63

AAC RH 6.8 0 231.0 1.00 0.00 N/A6 264.0 1.14 0.10 1.00

12 264.0 1.14 0.06 0.6218 259.0 1.12 0.08 0.7724 259.0 1.12 0.07 0.75

AAG RD 8.8 0 534.0 1.00 0.08 1.006 505.0 0.95 4.57 60.93

12 500.0 0.94 4.86 64.8018 495.0 0.93 4.87 64.9324 495.0 0.93 4.83 64.40

AAB RD 8.8 0 275.0 1.00 5.53 1.006 266.0 0.97 5.53 1.00

12 256.0 0.93 5.43 0.9818 268.0 0.97 5.09 0.9224 255.0 0.93 5.09 0.92

AAF RD 9.7 0 580.0 1.00 3.76 1.006 550.0 0.95 5.65 1.50

42 540.0 0.93 5.28 1.4018 540.0 0.93 5.09 1.3524 530.0 0.91 4.79 1.27

103

ECS Data

Air Cond ECS ECS Water Retained

Asph Agg Voids Time M r Mr Perm PermCode Code (%) (hr) (ksi) Ratio (E-3 cm/s) Ratio

AAM RD 10.4 0 430.0 1.00 0.19 1.006 402.0 0.93 2.76 14.76

12 380.0 0.88 2.31 12.3518 364.0 0.85 2.46 13.1624 390.0 0.91 2.46 13.16

AAC RJ 8.0 0 380.0 1.00 4.76 1.006 294.0 0.77 4.99 1.05

12 265.0 0.70 4.99 1.0518 260.0 0.68 4.69 0.99

24 254.0 0.67 4.40 0.92

AAD RH 6.9 0 230.0 1.00 0.00 N/A6 222.0 0.97 2.68 1.00

12 220.0 0.96 3.59 1.3418 219.0 0.95 2.72 1.0124 218.0 0.95 3.07 1.15

AAK RH 7.6 0 481.0 1.00 0.09 1.006 403.0 0.84 1.73 18.40

12 394.0 0.82 1.98 21.0618 373.0 0.78 1.75 18.6224 370.0 0.77 1.65 17.55

AAF RD 9.9 0 640.0 1.00 5.69 1.006 510.0 0.80 6.16 1.08

12 494.0 0.77 5.69 1.0018 488.0 0.76 5.69 1.0024 500.0 0.78 5.69 1.00

AAM RH 7.7 0 485.0 1.00 2.35 1.006 362.0 0.75 3.54 1.51

12 350.0 0.72 2.83 1.2018 348.0 0.72 2.73 I. 1624 345.0 0.71 2.59 1.10

104